# Salesbud API — full documentation

Generated from https://docs.salesbud.com.br. Every page, in reading order.

---

# Salesbud API

Source: https://docs.salesbud.com.br/

A meeting recorded by the bot and a call captured from VoIP are different
    resources, so they live at different paths: `/v1/meetings` and `/v1/calls`.
    Same shape, own identifiers.
  
  
    A credential belongs to a company, never to a user. Every read is filtered
    by that company — there is no parameter that widens it.
  
  
    Resources are addressed by opaque, prefixed ids (`mtg_`, `call_`, `usr_`).
    Internal sequential ids are never exposed.
  
  
    Every page is available as plain Markdown, plus [`llms.txt`](/llms.txt) and
    [`llms-full.txt`](/llms-full.txt) for tools that read documentation
    directly.
  

## Your first request

Exchange your credential for an access token, then read a page of meetings.

```bash title="First request"
# 1. Get an access token (valid for 1 hour)
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d "client_id=$SALESBUD_CLIENT_ID" \
  -d "client_secret=$SALESBUD_CLIENT_SECRET"

# 2. Read completed meetings from a period
curl "https://api.salesbud.com.br/v1/meetings?meeting_after=2026-01-01T00:00:00Z&limit=50" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

The response is a page of meetings plus a cursor:

```json title="200 OK"
{
  "data": [
    {
      "id": "mtg_H9_chV2YG6UGE0n31AqvDQ",
      "object": "meeting",
      "title": "Discovery — Acme",
      "status": "completed",
      "type": "video",
      "audience": "external",
      "meeting_at": "2026-01-14T13:00:00.000Z",
      "duration_seconds": 2840,
      "participants": [{ "email": "buyer@acme.com", "kind": "external" }],
      "owner": { "id": "usr_CKAyHd20jJb6GGJqhNF9vA", "email": "rep@yourcompany.com" },
      "transcript": { "available": true, "status": "completed" }
    }
  ],
  "pagination": { "limit": 50, "has_more": true, "next_cursor": "cur_..." }
}
```

## Where to go next

| If you want to… | Read |
| --- | --- |
| Make your first call end to end | [Quickstart](/get-started/quickstart/) |
| Understand credentials, tokens and scopes | [Authentication](/guides/authentication/) |
| Know when a record is a `meeting` and when it is a `call` | [Meetings and calls](/guides/meetings-and-calls/) |
| Walk a full history safely | [Pagination](/guides/pagination/) |
| Handle failures correctly | [Errors](/guides/errors/) |
| See every endpoint, field and status | [API reference](/reference/meetings/listmeetings/) |

:::note[Access is granted per company]
The API is gated by the `API_ACCESS` company feature. If your company does not
have it enabled, token issuance fails with `403 API_ACCESS_DISABLED` — talk to
your Salesbud contact rather than retrying.
:::

---

# The answer object

Source: https://docs.salesbud.com.br/objects/answer

Returned by the template answer routes, one entry per question of the record's default template.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque answer identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>meeting_answer</code>
  </div>

Resource type. Always meeting_answer, including for calls.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>object</code>
  </div>

The question this answers.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.id</code>
    <code>string</code>
  </div>

Opaque question identifier, stable across records that share the template. It is the field to key on, not the text, which the company can reword.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.text</code>
    <code>string</code>
  </div>

Question text as the template asks it, sanitized.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.type</code>
    <code>text</code> · <code>option</code> · <code>multi_select</code> · <code>boolean</code> · <code>number</code>
  </div>

What the template expects as an answer. The answer itself always arrives as a string, so this is what says how to parse it.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.order</code>
    <code>integer</code>
  </div>

Position of the question in the template, starting at zero.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answer</code>
    <code>string</code>
  </div>

Generated answer. A question that produced none is left out of the list rather than returned with a null answer.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
  </div>

When the answer last changed.

</div>

</div>

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# The call object

Source: https://docs.salesbud.com.br/objects/call

Returned by every route under `/v1/calls`. A call answers the same five routes as a meeting and carries the same fields except `bot_history`; what else changes is which records the collection holds and how they usually look.

Every field is documented on [the meeting object](/objects/meeting/). What follows is only what differs.

## What differs

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
  </div>

**`call`, never `meeting`.** It is the field that says which resource you hold.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
  </div>

**Prefixed `call_`.** A `mtg_` id does not resolve under /v1/calls, and the reverse is also a 404.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">participants</code>
  </div>

**Usually a single entry carrying `phone`.** A VoIP record identifies the other party by number, not by email.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">type</code>
  </div>

**Usually `audio`, but not always.** Calls in video exist; `object` is what tells the resource apart, never `type`.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">enablement.meeting_type</code>
  </div>

**Often `null`.** The classifier was built for sales meetings and does not label calls.

</div>

</div>

:::note[Fields a call does not have]
`bot_history` is not on this object: a VoIP capture has no recording bot, so the field never
applies and the response omits it rather than sending an empty array. A schema that validates a
call rejects the field.

`enablement.meeting_type` stays, and is often `null`. That null is legitimate — the sales-stage
classifier does not label calls today, but it could. It is not the same case as the bot.
:::

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# The evaluation object

Source: https://docs.salesbud.com.br/objects/evaluation

Returned by the overall evaluation routes. Every field except `object` and `status` is null until the evaluation completes.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
    nullable
  </div>

Opaque evaluation ID, null while no evaluation exists.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>overall_meeting_evaluation</code>
  </div>

Resource type. Always overall_meeting_evaluation, including for calls.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_id</code>
    <code>string</code>
  </div>

The record this evaluation belongs to. Carries the prefix of its collection: mtg_ or call_.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>not_started</code> · <code>processing</code> · <code>completed</code> · <code>failed</code>
  </div>

Public processing state.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">score</code>
    <code>integer</code>
    nullable
  </div>

Score from 0 to 10, or null until the evaluation completes.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">justification</code>
    <code>string</code>
    nullable
  </div>

Why the score was given, or null until the evaluation completes.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
    nullable
  </div>

When the evaluation was created, or null when none exists yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
    nullable
  </div>

When the evaluation last changed, or null when none exists yet.

</div>

</div>

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# The meeting object

Source: https://docs.salesbud.com.br/objects/meeting

Returned by every route under `/v1/meetings`. A call carries the same fields except `bot_history` — see [the call object](/objects/call/) for what differs.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Always prefixed `mtg_`. A `call_` id does not address a meeting: the API refuses it rather than resolving it in the other collection.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>meeting</code>
  </div>

Resource type. A record captured from a VoIP integration is a `call`; everything else is a `meeting`. Independent of `type`, which describes the media: a call can be video and a meeting can be audio.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">title</code>
    <code>string</code>
  </div>

Sanitized meeting title.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>completed</code>
  </div>

Only semantically completed meetings are public in V1.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">type</code>
    <code>video</code> · <code>audio</code>
  </div>

Meeting media type.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">audience</code>
    <code>internal</code> · <code>external</code>
  </div>

Meeting audience classification.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_at</code>
    <code>string date-time</code>
  </div>

Meeting instant in UTC.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">duration_seconds</code>
    <code>integer</code>
  </div>

Non-negative meeting duration.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">no_show</code>
    <code>boolean</code>
  </div>

Whether the completed meeting had no attendee.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">participants</code>
    array of <a href="#meeting-participant">object</a>
  </div>

Normalized external participants.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">owner</code>
    <a href="#meeting-owner">object</a>
  </div>

The user the record belongs to. Company scoping follows this user's default company.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">template</code>
    <a href="#template-reference">object</a>
    nullable
  </div>

Public meeting template reference when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">tags</code>
    array of <a href="#tag-reference">object</a>
  </div>

Public tag references.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">customer_questions</code>
    array of <a href="#meeting-customer-question">object</a>
  </div>

Customer questions detected in the call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">competitors</code>
    array of <a href="#meeting-competitor">object</a>
  </div>

Competitors mentioned in the call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">enablement</code>
    <a href="#meeting-enablement">object</a>
    nullable
  </div>

Enablement summary when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">accounts</code>
    array of <a href="#account-reference">object</a>
  </div>

Company-scoped accounts linked to the meeting.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">overall_evaluation</code>
    <code>object</code>
  </div>

Overall evaluation readiness and score summary.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill_scores</code>
    array of <a href="#meeting-skill-score">object</a>
  </div>

Per-skill scores; empty when no skill evaluation exists.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">transcript</code>
    <code>object</code>
  </div>

Transcript availability and its consistent processing state.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
  </div>

UTC instant when the meeting entered Salesbud.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
  </div>

UTC instant of the last change to the meeting record itself. Informational only: it does not advance when a dependent resource such as a transcript, tag or answer changes without touching the meeting row, so it is not a reliable incremental-sync marker.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">bot_history</code>
    array of <a href="#meeting-bot-event">object</a>
  </div>

Recording bot events. Present only on a meeting: a VoIP capture has no bot, so a call does not carry this field at all.

</div>

</div>

## Nested objects

### Meeting participant

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">display_name</code>
    <code>string</code>
    optional
  </div>

Sanitized external participant name when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    optional
  </div>

Normalized lowercase email when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">phone</code>
    <code>string</code>
    optional
  </div>

Normalized phone number when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">kind</code>
    <code>external</code>
  </div>

Participant classification in V1.

</div>

</div>

### Meeting owner

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque owner identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public owner display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    nullable
  </div>

Normalized owner email, or null when legacy data is malformed.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">team</code>
    <a href="#meeting-owner-team">object</a>
    nullable
  </div>

Owner default team inside the authenticated company.

</div>

</div>

### Meeting owner team

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque team identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Team display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">role</code>
    <code>participant</code> · <code>leadership</code> · <code>observer</code>
  </div>

Role of the owner in the team. `participant` sees only their own meetings; `leadership` sees the team's and is visible to it; `observer` sees the team's without appearing to the other members.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">is_default</code>
    <code>boolean</code>
  </div>

Whether this is the owner's default team in the company. False when the owner has no default team there and the API reported another one instead, which is why the field exists rather than being implied.

</div>

</div>

### Template reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque public template identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answers</code>
    array of <a href="#meeting-template-answer">object</a>
  </div>

Sanitized template questions and meeting answers.

</div>

</div>

### Meeting template answer

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">field</code>
    <code>string</code>
  </div>

Opaque public template question field identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>string</code>
  </div>

Question text as it was asked, sanitized.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answer</code>
    <code>string</code>
  </div>

Answer generated for this question. A question that produced none is left out of the list rather than returned with a null answer.

</div>

</div>

### Tag reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque public tag identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public display name.

</div>

</div>

### Meeting customer question

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>string</code>
  </div>

Question the customer asked during the conversation.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">category</code>
    <code>string</code>
  </div>

Category the question was classified into.

</div>

</div>

### Meeting competitor

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque competitor identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Competitor name as registered by the company, not as spoken.

</div>

</div>

### Meeting enablement

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">next_steps</code>
    <code>string</code>
    nullable
  </div>

Next steps agreed in the conversation, or null when none were identified.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_type</code>
    <code>qualification</code> · <code>negotiation</code> · <code>proposal</code> · <code>closing</code> · <code>technical_meeting</code> · <code>questions_clarifications</code>
    nullable
  </div>

Sales stage the conversation was classified as, or null. Often null on calls: the classifier was built for sales meetings.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration</code>
    <code>object</code>
  </div>

How long each side spoke.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users</code>
    array of <code>object</code>
  </div>

One entry per company user who spoke. A user who did not speak is left out rather than reported with zero.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users[].user</code>
    <a href="#public-user-reference">object</a>
  </div>

A company user as the API exposes them anywhere other than `owner`: the same identifier, without the team.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users[].duration_seconds</code>
    <code>integer</code>
  </div>

How long this user spoke, in seconds.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.others_seconds</code>
    <code>integer</code>
    nullable
  </div>

How long everyone outside the company spoke, in seconds, or null when the split was not measured. Not broken down per person: the other side is not identified by user.

</div>

</div>

### Public user reference

A company user as the API exposes them anywhere other than `owner`: the same identifier, without the team.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque user identifier, the same one `owner.id` carries.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public user display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    nullable
  </div>

Normalized lowercase user email, or null when legacy data is malformed.

</div>

</div>

### Account reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque account identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">domain</code>
    <code>string</code>
  </div>

Account email domain, lowercased.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">cnpj</code>
    <code>string</code>
    nullable
  </div>

Brazilian company registration number as stored, with or without punctuation, or null when unknown.

</div>

</div>

### Meeting skill score

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">user</code>
    <a href="#public-user-reference">object</a>
  </div>

The user the score refers to.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill</code>
    <code>object</code>
  </div>

The skill that was scored.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill.id</code>
    <code>string</code>
  </div>

Opaque skill identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill.name</code>
    <code>string</code>
  </div>

Skill name as the company registered it.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">score</code>
    <code>integer</code>
  </div>

Score from 0 to 10.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">justification</code>
    <code>string</code>
    nullable
  </div>

Why the score was given, or null when the evaluation produced none.

</div>

</div>

### Meeting bot event

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">code</code>
    <code>string</code>
    nullable
  </div>

What happened to the recording bot, as a stable code.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">subcode</code>
    <code>string</code>
    nullable
  </div>

Further detail on the event, or null when the code carries none.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">occurred_at</code>
    <code>string date-time</code>
  </div>

When the event happened.

</div>

</div>

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# The pagination object

Source: https://docs.salesbud.com.br/objects/pagination

Accompanies every list response. Drive the loop by `has_more`, never by the size of `data` — see [Pagination](/guides/pagination/).

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">limit</code>
    <code>integer</code>
  </div>

Number of resources requested.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">has_more</code>
    <code>boolean</code>
  </div>

Whether another page is available. A page can be short or empty while this is true, so always continue while it is true instead of relying on the size of data.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">next_cursor</code>
    <code>string</code>
    nullable
  </div>

Signed opaque cursor for the next page.

</div>

</div>

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# The transcript object

Source: https://docs.salesbud.com.br/objects/transcript

Returned by the transcript routes. An unavailable transcript is a successful response with `available: false`, not an error.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>transcript</code>
  </div>

Resource type. Always `transcript`, including for a call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_id</code>
    <code>string</code>
  </div>

The record this transcript belongs to. Carries the prefix of its collection: `mtg_` or `call_`.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">available</code>
    <code>true</code> · <code>false</code>
  </div>

Whether there is a transcript to read. `false` is a successful response carrying the reason in `status`, not an error.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>completed</code> · <code>not_started</code> · <code>processing</code> · <code>failed</code>
  </div>

Processing state. Always `completed` while `available` is true; otherwise why the transcript is not there yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">variant</code>
    <code>enhanced</code> · <code>original</code>
    nullable
  </div>

Which transcript was returned: `enhanced` when the improved version exists, `original` otherwise. Null while none is available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">utterances</code>
    array of <a href="#meeting-transcript-utterance">object</a>
  </div>

The transcript itself, in chronological order. Always empty while the transcript is unavailable.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
    nullable
  </div>

When the transcript was first stored, or null when none exists yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
    nullable
  </div>

When the transcript last changed, or null when none exists yet.

</div>

</div>

## Nested objects

### Meeting transcript utterance

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaker</code>
    <code>string</code>
    nullable
  </div>

Sanitized speaker label when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">text</code>
    <code>string</code>
  </div>

Sanitized utterance text.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">start_ms</code>
    <code>integer</code>
  </div>

Utterance start offset in milliseconds.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">end_ms</code>
    <code>integer</code>
  </div>

Utterance end offset in milliseconds.

</div>

</div>

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Quickstart

Source: https://docs.salesbud.com.br/get-started/quickstart

This walks from a fresh credential to a page of meetings. It takes two requests.

## Before you start

You need a **client ID** and **client secret**, issued by Salesbud for your
company. Credentials are not self-service: ask your Salesbud contact, who will
also confirm that the `API_ACCESS` feature is enabled for your company.

The secret is shown once, at creation. Store it in a secret manager — it cannot
be read back, only rotated.

1. **Get an access token**

   Tokens are issued through OAuth 2.0 client credentials. Send the credential
   as form-encoded fields:

   
   
   ```bash
   curl -X POST https://api.salesbud.com.br/oauth/token \
     -d 'grant_type=client_credentials' \
     -d "client_id=$SALESBUD_CLIENT_ID" \
     -d "client_secret=$SALESBUD_CLIENT_SECRET"
   ```
   
   
   ```js
   const response = await fetch("https://api.salesbud.com.br/oauth/token", {
     method: "POST",
     headers: { "Content-Type": "application/x-www-form-urlencoded" },
     body: new URLSearchParams({
       grant_type: "client_credentials",
       client_id: process.env.SALESBUD_CLIENT_ID,
       client_secret: process.env.SALESBUD_CLIENT_SECRET,
     }),
   });

   const { access_token, expires_in } = await response.json();
   ```
   
   
   ```python
   import os, requests

   response = requests.post(
       "https://api.salesbud.com.br/oauth/token",
       data={
           "grant_type": "client_credentials",
           "client_id": os.environ["SALESBUD_CLIENT_ID"],
           "client_secret": os.environ["SALESBUD_CLIENT_SECRET"],
       },
   )
   access_token = response.json()["access_token"]
   ```
   
   

   The endpoint requires `application/x-www-form-urlencoded`. Sending JSON
   returns `415 UNSUPPORTED_MEDIA_TYPE`.

   ```json title="200 OK"
   {
     "access_token": "eyJhbGciOiJSUzI1NiIs...",
     "token_type": "Bearer",
     "expires_in": 3600,
     "scope": "meetings.read calls.read"
   }
   ```

2. **Confirm what the token can do**

   `GET /v1/context` reports the company, the scopes and the effective rate
   limit behind the token. It is the fastest way to tell a configuration
   problem from an integration bug.

   ```bash
   curl https://api.salesbud.com.br/v1/context \
     -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
   ```

3. **Read a page of meetings**

   ```bash
   curl "https://api.salesbud.com.br/v1/meetings?meeting_after=2026-01-01T00:00:00Z&limit=50" \
     -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
   ```

   Only meetings your company owns and that finished processing are returned.
   See [Meetings and calls](/guides/meetings-and-calls/) for what "finished"
   means and why VoIP records live at `/v1/calls` instead.

4. **Walk the rest of the history**

   Pass `next_cursor` back unchanged, keeping every filter identical, and loop
   on `has_more` — never on the size of `data`.

   ```js
   let cursor = null;
   do {
     const url = new URL("https://api.salesbud.com.br/v1/meetings");
     url.searchParams.set("meeting_after", "2026-01-01T00:00:00Z");
     url.searchParams.set("limit", "50");
     if (cursor) url.searchParams.set("cursor", cursor);

     const page = await fetch(url, {
       headers: { Authorization: `Bearer ${accessToken}` },
     }).then((r) => r.json());

     for (const meeting of page.data) await handle(meeting);
     cursor = page.pagination.next_cursor;
   } while (cursor);
   ```

   A page can come back short — even empty — while `has_more` is `true`. That
   is expected: the service caps how much it scans per request. [Pagination](/guides/pagination/)
   explains the guarantees.

## Renewing the token

There is no refresh token. Client credentials is a machine-to-machine grant with
no user to re-consent, so [RFC 6749 §4.4.3](https://www.rfc-editor.org/rfc/rfc6749#section-4.4.3)
says not to issue one. When the token expires, call `/oauth/token` again with
the same credential.

In practice: cache the token for slightly less than `expires_in`, and re-issue on
`401 INVALID_ACCESS_TOKEN` rather than on a timer alone — a credential can be
revoked before the clock runs out.

## If a model is the one reading

If the consumer is an agent rather than a service, you can skip the HTTP client
entirely: the [MCP server](/guides/mcp/) exposes these same operations as tools
for Claude Desktop, Claude Code and Cursor, and takes care of token renewal and
retries. You still need the credential above.

## Common first-call failures

| Response | What it means |
| --- | --- |
| `415 UNSUPPORTED_MEDIA_TYPE` | The token request was sent as JSON. Use form encoding. |
| `401 INVALID_CLIENT` | Wrong client ID or secret, or the credential was revoked. |
| `403 API_ACCESS_DISABLED` | The company does not have the `API_ACCESS` feature. Retrying will not help. |
| `403 INSUFFICIENT_SCOPE` | The token lacks a scope the route requires — `calls.read` for calls, and `transcriptions.read` on top of the collection scope for transcripts. |
| `404 RESOURCE_NOT_FOUND` | The id does not exist, belongs to another company, or belongs to the other collection. |

---

# Authentication

Source: https://docs.salesbud.com.br/guides/authentication

Every request carries a bearer access token, obtained by exchanging a client
credential at `POST /oauth/token`.

## A credential is a company, not a user

The credential identifies **your company**. It is not tied to a person, a login,
a team or a selected workspace, and it does not inherit anybody's permissions.
Everything the token can read is scoped to the company it was issued for, and no
parameter widens that.

This is why there is no user login flow: there is no user.

## Getting a token

```bash
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d "client_id=$SALESBUD_CLIENT_ID" \
  -d "client_secret=$SALESBUD_CLIENT_SECRET"
```

The endpoint accepts only `application/x-www-form-urlencoded`. Credentials may
also be sent with HTTP Basic instead of in the body.

```json title="200 OK"
{
  "access_token": "eyJhbGciOiJSUzI1NiIs...",
  "token_type": "Bearer",
  "expires_in": 3600,
  "scope": "meetings.read calls.read transcriptions.read"
}
```

Tokens are signed RS256 and carry the company, the granted scopes and the
credential version.

## Scopes

| Scope | Grants |
| --- | --- |
| `meetings.read` | List and retrieve meetings, their template answers and their overall evaluations. |
| `calls.read` | List and retrieve calls, their template answers and their overall evaluations. |
| `transcriptions.read` | Read transcripts, alongside the scope of the collection the transcript belongs to. Requested separately because a transcript is the most sensitive payload the API serves. |

A transcript therefore needs two scopes: `meetings.read` and
`transcriptions.read` for a meeting transcript, `calls.read` and
`transcriptions.read` for a call transcript. `transcriptions.read` on its own
reads nothing.

A token receives the scopes attached to its credential. Requesting a route
outside them returns `403 INSUFFICIENT_SCOPE` — that is the gate working, not a
missing route.

Call `GET /v1/context` to see the scopes a token actually carries.

## Lifetime and renewal

Tokens expire after `expires_in` seconds (3600 by default).

**There is no refresh token, by design.** Client credentials is a
machine-to-machine grant with no user to re-consent, so
[RFC 6749 §4.4.3](https://www.rfc-editor.org/rfc/rfc6749#section-4.4.3) says not
to issue one — a refresh token would be a second long-lived secret with the same
power as the first. Renewal is calling `/oauth/token` again.

Cache the token slightly under `expires_in`, and also re-issue when a request
returns `401 INVALID_ACCESS_TOKEN`. A credential can be revoked before the clock
runs out; a timer alone will not notice.

## Rotating a secret

Rotation issues a new secret while the previous one keeps working for **seven
days**, so you can deploy the new value without a window of failed requests.

At most two secrets are valid at once. Revoke the expiring one before starting
another rotation, otherwise the second rotation is refused with
`409 ROTATION_IN_PROGRESS`.

Revoking or disabling the whole client invalidates issued tokens immediately —
that is the lever to pull on a suspected leak, not rotation.

## IP allowlists

A credential can carry an optional list of IPv4/IPv6 CIDRs. When set, it is
enforced both when issuing a token and on every request that uses it, so a
leaked token is useless from outside your network.

:::caution[The secret is shown once]
It is stored only as an HMAC-SHA256 hash with a service-side pepper and cannot be
read back. If it is lost, rotate — there is no recovery path.
:::

---

# Errors

Source: https://docs.salesbud.com.br/guides/errors

Errors use one envelope across the API:

```json
{
  "error": {
    "type": "ForbiddenError",
    "title": "Forbidden",
    "detail": "The transcriptions.read scope is required.",
    "code": "INSUFFICIENT_SCOPE",
    "request_id": "req_9bfbe80ac1f24e0a9d5b1e5e7a0f2c11"
  }
}
```

## Branch on `code`

`code` is the stable, machine-readable field. It does not change with wording,
translation or refactoring.

`title` and `detail` are human text and **may be reworded** between releases.
`type` is the error class and is stable, but coarser than `code` — several codes
share one class. Never parse `detail`.

Quote `request_id` when contacting support: it identifies the exact request in
our logs.

## Codes

| HTTP | `code` | Meaning |
| --- | --- | --- |
| 400 | `INVALID_LIMIT` | Page size outside 1–100. |
| 400 | `INVALID_DATETIME` | A timestamp is not RFC 3339 with an explicit offset. |
| 400 | `INVALID_FILTER` | A filter value or combination is unsupported. |
| 400 | `INVALID_CURSOR` | The cursor was edited, or the filters changed mid-traversal. |
| 400 | `INVALID_REQUEST` | Request validation failed, including unknown query parameters. |
| 401 | `MISSING_ACCESS_TOKEN` | No bearer token was sent. |
| 401 | `INVALID_ACCESS_TOKEN` | The token is invalid, expired, or its credential was revoked. |
| 403 | `INSUFFICIENT_SCOPE` | The token does not carry the scope the route requires. |
| 403 | `API_ACCESS_DISABLED` | The company does not have API access enabled. |
| 404 | `RESOURCE_NOT_FOUND` | Absent, owned by another company, or addressed in the wrong collection. |
| 413 | `PAYLOAD_TOO_LARGE` | The request body exceeds the accepted size. |
| 415 | `UNSUPPORTED_MEDIA_TYPE` | Wrong content type — most often JSON sent to `/oauth/token`. |
| 429 | `RATE_LIMIT_EXCEEDED` | A rate-limit policy rejected the request. |
| 503 | `SERVICE_UNAVAILABLE` | A fail-closed dependency is unavailable. Retry with backoff. |
| 503 | `AUDIT_UNAVAILABLE` | A successful read could not be recorded in the audit trail, so it was not returned. |

## Unknown identifiers are always 404

An id belonging to another company returns `404 RESOURCE_NOT_FOUND`, not `403`.
The API does not confirm that a resource exists somewhere else — a `403` would
leak that fact to anyone probing ids.

## The token endpoint is different

`POST /oauth/token` returns the OAuth 2.0 error shape that
[RFC 6749 §5.2](https://www.rfc-editor.org/rfc/rfc6749#section-5.2) mandates, so
standard OAuth clients keep working:

```json
{
  "error": "invalid_client",
  "error_description": "Client authentication failed."
}
```

## What is never in an error

Stack traces, source paths, tokens, secrets, request bodies and domain payloads
are never returned. An unexpected failure returns a fixed generic message with a
`request_id`; the detail stays in our logs.

## Retrying

| Status | Retry? |
| --- | --- |
| `400`, `403`, `404`, `413`, `415` | No. The request will fail identically. |
| `401` | Once, after re-issuing the token. |
| `429` | Yes, after `Retry-After`. |
| `503` | Yes, with exponential backoff and jitter. |

---

# MCP server

Source: https://docs.salesbud.com.br/guides/mcp

This API serves a [Model Context Protocol](https://modelcontextprotocol.io)
server over HTTP at `https://api.salesbud.com.br/mcp`. Registered once as a custom
connector in claude.ai, it lets the model read meetings, calls, transcripts,
template answers, evaluations, email and WhatsApp conversations with your
credential — nothing to install.

It is a client of the public API, not a new surface. Same OAuth credential, same
scopes, same rate limits, **read operations only** — nothing it can reach is
anything the HTTP API would not return to the same credential.

## Before you start

You need an API credential with the scopes you intend to use, the `API_ACCESS`
feature enabled for your company, and owner access to the Claude organization
that will hold the connector. Credentials are not self-service; see the
[Quickstart](/get-started/quickstart/).

## Add the connector

An owner of the Claude organization adds a custom connector pointing at
`https://api.salesbud.com.br/mcp`. Authentication is **No sign-in** — this server has no
OAuth flow — and the credential of the company whose data the connector should
read travels in two request headers:

| Header            | Value           |
| ----------------- | --------------- |
| `X-Client-Id`     | `sb_client_...` |
| `X-Client-Secret` | `sb_secret_...` |

Any other HTTP client may send the same credential as RFC 7617 Basic instead:

```
Authorization: Basic <base64 of client_id:client_secret>
```

Two consequences of a credential entered once by an administrator:

- **It is stored on Anthropic's side and shared by the whole Claude
  organization.** Everyone in it reads whatever the scopes allow, and the audit
  trail records the credential, not the person who asked.
- **Revoking means rotating.** Rotating the secret breaks every other
  integration using the same credential, so give the connector one of its own.

If the credential restricts access by IP, the allowlist has to include
Anthropic's egress range, `160.79.104.0/21`: the request reaches this API from
there, not from the user's browser.

## Tools

One tool per public read operation. The scope column is what the credential has
to carry — a missing scope comes back as a readable `INSUFFICIENT_SCOPE` message
the model can report, not an exception.

| Tool                             | Scope                                     |
| -------------------------------- | ----------------------------------------- |
| `get_api_context`                | none beyond a valid credential            |
| `list_meetings`                  | `meetings.read`                           |
| `get_meeting`                    | `meetings.read`                           |
| `get_meeting_overall_evaluation` | `meetings.read`                           |
| `get_meeting_answers`            | `meetings.read`                           |
| `get_meeting_transcript`         | `meetings.read` + `transcriptions.read`   |
| `list_calls`                     | `calls.read`                              |
| `get_call`                       | `calls.read`                              |
| `get_call_overall_evaluation`    | `calls.read`                              |
| `get_call_answers`               | `calls.read`                              |
| `get_call_transcript`            | `calls.read` + `transcriptions.read`      |
| `list_emails`                    | `emails.read`                             |
| `get_email`                      | `emails.read`                             |
| `list_email_messages`            | `emails.read` + `emails.content.read`     |
| `list_whatsapp_conversations`    | `whatsapp.read`                           |
| `get_whatsapp_conversation`      | `whatsapp.read`                           |
| `list_whatsapp_messages`         | `whatsapp.read` + `whatsapp.content.read` |

Ask `get_api_context` first when you are unsure what a credential reaches: it
answers with the client, the company, the scopes and the per-minute limit.

`/oauth/token` is deliberately not a tool — the server issues and renews the
token itself, and exposing it would burn the OAuth bucket and put the token in
the conversation. The liveness and readiness probes are not tools either.

## Transcripts arrive in windows

A one-hour transcript is thousands of utterances and does not belong in a
model's context in one piece. The transcript tools return 200 utterances per
call and say how many are left; the model passes `offset` to continue. The API
returns the transcript whole — this windowing lives in the MCP server.

A record with no transcript is not an error: the tool reports
`transcript unavailable (status: not_started)`, which is the contract described
in [Meetings and calls](/guides/meetings-and-calls/).

## Email bodies cost more than email metadata

`list_emails` and `get_email` return conversation metadata — participants, connected
mailboxes, linked accounts, counts — and never a body. Reading the messages is a
separate tool, `list_email_messages`, and it asks for more: `emails.content.read` on
top of `emails.read`, plus its own rate-limit policy, tighter than the rest of the
API. Ask for a page at a time instead of walking a whole mailbox.

A body longer than 2,000 characters is trimmed in the text the model reads and kept
whole in `structuredContent`. Attachment content, HTML bodies and `bcc` are never
served by the API, so no tool reaches them.

## WhatsApp text costs more than WhatsApp metadata

`list_whatsapp_conversations` and `get_whatsapp_conversation` return conversation
metadata — contact, seller, linked accounts, activity instants — and never a message.
Reading the messages is a separate tool, `list_whatsapp_messages`, and it asks for more:
`whatsapp.content.read` on top of `whatsapp.read`, plus its own rate-limit policy. Ask
for a page at a time instead of walking a whole history.

A text or an audio transcription longer than 2,000 characters is trimmed in the text the
model reads and kept whole in `structuredContent`. A message deleted for everyone stays as
a tombstone with no content; an edited message shows its current text. Media content,
provider identifiers, reactions and read state never reach the model — the API does not
serve them. See the [WhatsApp reference](/reference/whatsapp/listwhatsappconversations/).

## What the server handles, and what it does not

It handles token renewal with a safety margin, one retry on `401` with a fresh
token, `Retry-After` on `429`, backoff on `503`, and immediate failure on the
other `4xx`. Errors reach the model carrying `code` and `request_id`, so a bad
date filter comes back as `INVALID_DATETIME` and the model corrects the format
itself.

It does **not** paginate for you. The model follows `has_more` and `next_cursor`
like any other client, and the tool description says so, because
[a short page is not the last page](/guides/pagination/).

It also does not give you incremental sync, because v1 has none: window by
`meeting_at` or `created_at` and reprocess idempotently by `id`.

<Aside type="caution" title="Give the connector its own credential">
  A company can hold 20 active clients. Issue one for the connector alone, with
  only the scopes it needs, so revoking it on the API credentials screen never
  disturbs your other integrations.
</Aside>

## Shared quota

Every credential of a company draws from the same 600 requests/minute company
bucket. An agent walking a long history consumes quota your other integrations
are using; see [Rate limits](/guides/rate-limits/).

---

# Meetings and calls

Source: https://docs.salesbud.com.br/guides/meetings-and-calls

A recording captured by the meeting bot and a recording captured from a VoIP
integration are different things in the product, so they are different resources
in the API.

| | Meetings | Calls |
| --- | --- | --- |
| Collection | `/v1/meetings` | `/v1/calls` |
| `object` | `meeting` | `call` |
| Identifier prefix | `mtg_` | `call_` |
| Where it comes from | Meeting bot, platform upload, mobile app | VoIP integration |

Both support the same five routes — list, retrieve, transcript, template answers
and overall evaluation — and carry the same fields except `bot_history`, which
only a meeting has. What else changes is the collection and the identifier
prefix.

## Do not infer the kind from the media type

`object` and `type` are independent axes:

- `object` says **what the record is** — `meeting` or `call`.
- `type` says **how it was captured** — `video` or `audio`.

A call can be video, and a meeting can be audio. In one production company, 3 of
49 calls are video and 6 of 863 meetings are audio. Branching on `type` to guess
the kind gets those wrong.

```json
{ "id": "call_IGmhT13We_ltOPvjNSCHKw", "object": "call", "type": "audio" }
{ "id": "call_J781tlst89klU_mWPMXgDA", "object": "call", "type": "video" }
```

## Identifiers do not cross collections

An id addresses a resource in exactly one collection. Asking the wrong one is a
`404`, not a redirect:

```bash
curl ".../v1/calls/mtg_H9_chV2YG6UGE0n31AqvDQ" -H "Authorization: Bearer $TOKEN"
# 404 RESOURCE_NOT_FOUND
```

This is deliberate: it keeps an id from silently resolving to a resource of a
different kind, and it makes a mis-routed request obvious in your logs instead of
returning something plausible.

## What is visible

A record appears in the API only when it is **completed** — the processing that
produces transcripts, answers and evaluations has finished. Records still
recording, processing or failed are not returned at all, rather than exposed with
a partial `status`. `status` is therefore always `completed` in v1.

Ownership follows the meeting owner's **default company**. If a user's default
company changes, their historical records move with them. This is a known
trade-off of resolving ownership at read time; see [Security](/guides/security/).

## Choosing a collection

Read `/v1/meetings` for sales conversations captured by the bot or uploaded, and
`/v1/calls` for phone conversations captured through VoIP. If you want both, page
each collection independently — there is no combined endpoint, because the two
have different volumes and different retention in practice.

:::tip[Fields that behave differently on calls]
A call does not carry `bot_history`: there is no recording bot in a VoIP capture, so the field is
omitted rather than empty — see [the call object](/objects/call/).

`participants` on a call is usually a single entry carrying `phone` rather than
`email`, because a VoIP record identifies the other party by number.
`enablement.meeting_type` is often `null` on calls: the classifier was trained on
sales meetings and does not label calls.
:::

---

# Pagination

Source: https://docs.salesbud.com.br/guides/pagination

Lists are paginated with signed keyset cursors ordered by ascending id. The
default page size is 50 and the maximum is 100.

```json
{
  "data": [ /* … */ ],
  "pagination": { "limit": 50, "has_more": true, "next_cursor": "cur_..." }
}
```

## Loop on `has_more`, never on `data.length`

This is the single most common integration bug against this API.

A page can come back **short, or even empty, while `has_more` is `true`**. The
service caps how much it scans per request, so when a window is sparse the
remaining rows are served on the next call rather than in a slower single
response.

```js
// Correct
let cursor = null;
do {
  const page = await fetchPage(cursor);
  await handle(page.data);
  cursor = page.pagination.next_cursor;
} while (cursor);
```

```js
// Wrong — stops early on a sparse window
let page = await fetchPage();
while (page.data.length > 0) { /* … */ }
```

## The cursor binds your filters

`next_cursor` is signed and covers every filter of the request that produced it.
Send it back **unchanged**, with **identical filters**. Changing a filter mid-walk,
or editing the cursor, returns `400 INVALID_CURSOR` instead of silently returning
a different slice.

To change filters, start a new traversal from no cursor.

## What ordering by id guarantees

Ordering by ascending id makes a traversal stable without a snapshot: records
created while you paginate get higher ids, so they land after your current
position and never shift a page you already read.

The trade-off is the mirror image: a record that becomes **completed** after you
have already passed its id is not picked up by the traversal in progress. It
appears on your next pass over that window.

This is why v1 has no incremental sync — there is no dedicated "became visible"
timestamp to make a delta honest. `updated_at` is informative only; do not use it
as a sync watermark. Re-query by period instead:

```bash
curl ".../v1/meetings?meeting_after=2026-01-01T00:00:00Z&meeting_before=2026-02-01T00:00:00Z"
```

## Filters

| Parameter | Filters on |
| --- | --- |
| `meeting_after` / `meeting_before` | When the meeting or call happened. The filter most integrations want. |
| `created_after` / `created_before` | When the record was created in Salesbud. |
| `owner_email` | Exact, case-insensitive owner match. |
| `type` | `video` or `audio` — the media, not the resource kind. |
| `audience` | `internal` or `external`. |
| `has_transcript` | Whether a transcript resource exists. |

All timestamps are RFC 3339 with an explicit offset. `2026-01-01T00:00:00Z` is
valid; `2026-01-01` is not, and returns `400 INVALID_DATETIME`.

---

# Rate limits

Source: https://docs.salesbud.com.br/guides/rate-limits

Requests are metered with distributed token buckets. Several apply at once, and
the strictest one wins.

| Bucket | Applies to | Quota |
| --- | --- | --- |
| `client` | Every data-plane request, per credential | 120/min, adjustable per credential |
| `company` | Every data-plane request, across all credentials of the company | 600/min |
| `transcript` | Transcript reads only, on top of the two above | 60/min |
| `oauth-client` | `POST /oauth/token`, per client id | 10/min |
| `oauth-ip` | `POST /oauth/token`, per source IP | 30/min |

A company may hold up to 20 active credentials, and together they never exceed
the company quota. Call `GET /v1/context` to read the credential quota actually
in effect for yours.

## Headers

Every response carries the current state. `RateLimit-Policy` lists every bucket
the request consumed; `RateLimit` reports the single one closest to exhaustion:

```http
RateLimit-Policy: "client";q=120;w=60, "company";q=600;w=60
RateLimit: "client";r=118;t=1
```

`q` is the quota, `w` the window in seconds, `r` what is left and `t` the seconds
until another token. `RateLimit` names whichever bucket is holding the fewest
tokens: normally your credential's, the transcript one on transcript reads, and
the company one once sibling credentials are spending it. Pace against the
ceilings in `RateLimit-Policy` — the bucket named in one reading is not the only
one binding you.

`r` counts tokens in the bucket, whose capacity is the burst rather than the
quota, so it does not start at `q` and count down. Treat it as room available
right now.

A rejected request returns `429 RATE_LIMIT_EXCEEDED` with `Retry-After` in
seconds. Wait that long — do not retry immediately, and add jitter so a fleet of
workers does not resynchronise into the next window.

## Behaviour when the limiter is degraded

Read paths **fail open**: if the limiter is briefly unreachable, business reads
keep serving rather than going dark.

Token issuance, transcript reads and internal control-plane routes **fail
closed** with `503 SERVICE_UNAVAILABLE`. These are the paths where serving
unmetered traffic is worse than serving none.

## The WAF limit is separate

A coarse per-IP limit runs at the WAF, before traffic reaches the service, and it
is not part of the quotas above: roughly 2000 requests per five minutes per
source IP, and 100 per five minutes on `/oauth/token`.

It fails differently, which matters when you meet it:

- it blocks with `403`, not `429`, so it reads as an authorization failure;
- there is no error code, no `Retry-After` and no rate limit headers, because the
  request never reaches the service;
- it counts by source IP, so every integration behind the same egress address
  shares it, whatever credentials they use.

A `403` with none of our error body, especially on `/oauth/token`, is this limit
rather than a missing scope.

## Staying under the limit

- Page with `limit=100` instead of many small pages — one request returns up to
  100 records at the same cost as one returning 10.
- Filter by period rather than walking the whole history repeatedly.
- Cache the access token; every `/oauth/token` call consumes its own bucket.
- Fetch transcripts only for records you will actually process. `has_transcript`
  lets you filter the list first.

---

# Security

Source: https://docs.salesbud.com.br/guides/security

## Company isolation

Every read is filtered by the company on the access token. The company comes
from the token claim, never from a request parameter, so there is no input that
widens the scope of a credential.

Resources are addressed by opaque, prefixed identifiers (`mtg_`, `call_`,
`usr_`, `tpl_`) that are allocated per company. Internal sequential ids are never
exposed, so record volume cannot be inferred and ids cannot be enumerated.

An identifier your company does not own returns `404`, never `403` — the API does
not confirm that a resource exists elsewhere.

## Ownership is resolved when you read

A record belongs to the **default company of the user who owns it**, evaluated at
query time rather than frozen when the record was created.

The consequence is worth stating plainly: **if a user's default company changes,
their historical records move with them.** A rep who moves between companies in
the same group takes their meeting history to the new company's API.

This is a deliberate trade-off of not denormalising ownership onto the record. If
that behaviour does not fit your data agreement, raise it with your Salesbud
contact before building on it.

## Transcripts are gated separately

`transcriptions.read` is a distinct scope from `meetings.read` and `calls.read`,
and a credential can hold the collection scopes without it. A transcript is the
most sensitive payload the API serves — verbatim customer speech — so access to
the metadata of a meeting does not imply access to what was said in it.

The gate composes rather than replacing the collection scope: a transcript
request is checked against both, so a credential limited to calls cannot reach a
meeting transcript by holding `transcriptions.read`.

Transcript text is sanitised before it leaves the service.

## Everything is audited

Every business read is recorded: which credential, which company, which route,
which resource ids, and the calling identity.

The audit write is **fail closed**. If the trail cannot be written, the read is
not returned — you get `503 AUDIT_UNAVAILABLE` instead. A successful response
therefore means the access was recorded.

## Credential hygiene

- Secrets are stored only as HMAC-SHA256 hashes with a service-side pepper. They
  cannot be read back, only rotated.
- Rotation keeps the previous secret valid for seven days so you can deploy
  without downtime; revoke it once the new one is live.
- On a suspected leak, **revoke the client** — that invalidates issued tokens
  immediately. Rotation does not, because the old secret stays valid.
- Set an IP allowlist when your integration runs from fixed addresses. It is
  enforced at token issuance and on every request, so a leaked token is useless
  from elsewhere.

## What never leaves the service

Stack traces, source paths, internal ids, tokens, secrets and request bodies are
never returned in a response, and never written to logs.

---

# Issue an access token

Source: https://docs.salesbud.com.br/reference/authentication/issueaccesstoken

post <code>{"/oauth/token"}</code>

### Authorization

No authentication.

### Example request

```bash
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d 'client_id=sb_client_...' \
  -d 'client_secret=sb_secret_...'
```

### Example response

```json title="200 OK"
{
  "access_token": "eyJhbGciOiJSUzI1NiJ9.EXAMPLE_PAYLOAD.EXAMPLE_SIGNATURE",
  "token_type": "Bearer",
  "expires_in": 3600,
  "scope": "meetings.read calls.read transcriptions.read"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Access token issued. |
| `400` | OAuth client-credentials request rejected. |
| `401` | OAuth client authentication failed. |
| `413` | The request body is larger than the accepted limit. |
| `415` | The request media type is not supported by this endpoint. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get one completed call

Source: https://docs.salesbud.com.br/reference/calls/getcall

get <code>{"/v1/calls/{call_id}"}</code>

Returns a single [call object](/objects/call/).

### Authorization

Requires the `calls.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `call_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": "call_ExampleCall01",
    "object": "call",
    "title": "Call to +55 11 99999-0000",
    "status": "completed",
    "type": "audio",
    "audience": "external",
    "meeting_at": "2026-01-15T18:28:01.000Z",
    "duration_seconds": 714,
    "no_show": false,
    "participants": [
      {
        "phone": "+5511999990000",
        "kind": "external"
      }
    ],
    "owner": {
      "id": "usr_ExampleOwner01",
      "name": "Example Rep",
      "email": "rep@example.com",
      "team": {
        "id": "team_ExampleTeam01",
        "name": "Inside Sales",
        "role": "participant",
        "is_default": true
      }
    },
    "template": {
      "id": "tpl_ExampleTemplate01",
      "name": "Discovery script",
      "answers": [
        {
          "field": "qst_ExampleQuestion01",
          "question": "What pains did the customer describe?",
          "answer": "Wants a demo before renewing the current contract."
        }
      ]
    },
    "tags": [],
    "customer_questions": [],
    "competitors": [],
    "enablement": {
      "next_steps": "Online meeting booked for Thursday at 13:30; send the invite today.",
      "meeting_type": null,
      "speaking_duration": {
        "users": [],
        "others_seconds": 677
      }
    },
    "accounts": [],
    "overall_evaluation": {
      "status": "not_started",
      "justification": null
    },
    "skill_scores": [],
    "transcript": {
      "available": false,
      "status": "not_started"
    },
    "created_at": "2026-01-15T19:22:08.000Z",
    "updated_at": "2026-01-15T19:22:53.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped call. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List the default template answers for a call

Source: https://docs.salesbud.com.br/reference/calls/getcallanswers

get <code>{"/v1/calls/{call_id}/answers"}</code>

Returns a list of [answer objects](/objects/answer/).

### Authorization

Requires the `calls.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `call_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/answers" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "ans_ExampleAnswer01",
      "object": "meeting_answer",
      "question": {
        "id": "qst_ExampleQuestion01",
        "text": "What pains did the customer describe?",
        "type": "text",
        "order": 1
      },
      "answer": "Manual CRM entry after every call, about two hours a day.",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Answers generated for the call owner context. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get a call transcript

Source: https://docs.salesbud.com.br/reference/calls/getcalltranscript

get <code>{"/v1/calls/{call_id}/transcript"}</code>

Returns the enhanced transcript when available, otherwise the original transcript. An unavailable transcript is represented by a successful response with available set to false and its current processing status.

Returns a single [transcript object](/objects/transcript/).

### Authorization

Requires the `calls.read` and `transcriptions.read` scopes together.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `call_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/transcript" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "object": "transcript",
    "meeting_id": "call_ExampleCall01",
    "available": false,
    "status": "processing",
    "variant": null,
    "utterances": [],
    "created_at": null,
    "updated_at": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped call transcript. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get the overall call evaluation

Source: https://docs.salesbud.com.br/reference/calls/getoverallcallevaluation

get <code>{"/v1/calls/{call_id}/evaluations/overall"}</code>

Returns a single [evaluation object](/objects/evaluation/).

### Authorization

Requires the `calls.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `call_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/evaluations/overall" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": null,
    "object": "overall_meeting_evaluation",
    "meeting_id": "call_ExampleCall01",
    "status": "not_started",
    "score": null,
    "justification": null,
    "created_at": null,
    "updated_at": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Overall evaluation state and result. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List completed calls

Source: https://docs.salesbud.com.br/reference/calls/listcalls

get <code>{"/v1/calls"}</code>

Calls are the records captured from a VoIP integration, addressed by their own `call_` identifiers. They answer the same five routes as meetings and carry the same fields except `bot_history`: a VoIP capture has no recording bot, so the field is absent, not empty.

Returns a list of [call objects](/objects/call/).

### Authorization

Requires the `calls.read` scope.

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |
| `created_after` | `string date-time` | No | Return records created after this instant. |
| `created_before` | `string date-time` | No | Return records created before this instant. |
| `meeting_after` | `string date-time` | No | Return records whose event instant is after this value. |
| `meeting_before` | `string date-time` | No | Return records whose event instant is before this value. |
| `owner_email` | `string email` | No | Exact case-insensitive owner email match. |
| `type` | `video` · `audio` | No | Media the record was captured as. Independent of the resource kind: a call can be video and a meeting can be audio. |
| `audience` | `internal` · `external` | No | Whether the record had participants from outside the company (external) or only internal ones. |
| `has_transcript` | `boolean` | No | Filter by existence of a transcript resource. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/calls?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "call_ExampleCall01",
      "object": "call",
      "title": "Call to +55 11 99999-0000",
      "status": "completed",
      "type": "audio",
      "audience": "external",
      "meeting_at": "2026-01-15T18:28:01.000Z",
      "duration_seconds": 714,
      "no_show": false,
      "participants": [
        {
          "phone": "+5511999990000",
          "kind": "external"
        }
      ],
      "owner": {
        "id": "usr_ExampleOwner01",
        "name": "Example Rep",
        "email": "rep@example.com",
        "team": {
          "id": "team_ExampleTeam01",
          "name": "Inside Sales",
          "role": "participant",
          "is_default": true
        }
      },
      "template": {
        "id": "tpl_ExampleTemplate01",
        "name": "Discovery script",
        "answers": [
          {
            "field": "qst_ExampleQuestion01",
            "question": "What pains did the customer describe?",
            "answer": "Wants a demo before renewing the current contract."
          }
        ]
      },
      "tags": [],
      "customer_questions": [],
      "competitors": [],
      "enablement": {
        "next_steps": "Online meeting booked for Thursday at 13:30; send the invite today.",
        "meeting_type": null,
        "speaking_duration": {
          "users": [],
          "others_seconds": 677
        }
      },
      "accounts": [],
      "overall_evaluation": {
        "status": "not_started",
        "justification": null
      },
      "skill_scores": [],
      "transcript": {
        "available": false,
        "status": "not_started"
      },
      "created_at": "2026-01-15T19:22:08.000Z",
      "updated_at": "2026-01-15T19:22:53.000Z"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor02"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped call page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get the authenticated integration context

Source: https://docs.salesbud.com.br/reference/context/getapicontext

get <code>{"/v1/context"}</code>

### Authorization

No authentication.

### Example request

```bash
curl "https://api.salesbud.com.br/v1/context" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "object": "api_context",
    "client": {
      "id": "sb_client_ExampleClient01",
      "name": "acme-sync"
    },
    "company": {
      "id": "cmp_ExampleCompany01",
      "name": "Example Company"
    },
    "scopes": [
      "meetings.read",
      "calls.read",
      "transcriptions.read"
    ],
    "rate_limit": {
      "requests_per_minute": 120
    }
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | API client, company, scopes and rate-limit context. |
| `401` | Missing or invalid credentials. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get an email conversation

Source: https://docs.salesbud.com.br/reference/emails/getemail

get <code>{"/v1/emails/{email_id}"}</code>

Conversation metadata only. An `eml_` id issued to another company, or any id with a different prefix, is a 404, not a hint.

### Authorization

Requires the `emails.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `email_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/emails/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": "eml_01JEXAMPLE",
    "object": "email_conversation",
    "subject": "Assunto de exemplo",
    "first_message_at": "2026-01-10T12:00:00.000Z",
    "last_message_at": "2026-01-12T15:30:00.000Z",
    "message_count": 3,
    "has_attachments": true,
    "last_message_direction": "inbound",
    "participants": [
      {
        "address": "vendedora@example.com",
        "name": "Vendedora Exemplo",
        "internal": true
      },
      {
        "address": "cliente@example.net",
        "name": "Cliente Exemplo",
        "internal": false
      }
    ],
    "mailboxes": [
      {
        "address": "vendedora@example.com",
        "owner": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        }
      },
      {
        "address": "vendas@example.com",
        "owner": null
      }
    ],
    "accounts": [
      {
        "id": "acc_01JEXAMPLE",
        "domain": "example.net",
        "cnpj": "00.000.000/0001-00"
      }
    ]
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped email conversation. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List the messages of an email conversation

Source: https://docs.salesbud.com.br/reference/emails/listemailmessages

get <code>{"/v1/emails/{email_id}/messages"}</code>

Messages in the order they were sent, oldest first, with plain-text bodies and attachment metadata (no attachment content). Governed by the `email_content` rate-limit policy in addition to the client and company policies. The cursor is bound to the conversation.

### Authorization

Requires the `emails.read` and `emails.content.read` scopes together.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `email_id` | `string` | Yes | — |

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/emails/mtg_01JEXAMPLE/messages?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "emsg_01JEXAMPLE",
      "object": "email_message",
      "conversation_id": "eml_01JEXAMPLE",
      "sent_at": "2026-01-12T15:30:00.000Z",
      "direction": "inbound",
      "subject": "Re: Assunto de exemplo",
      "snippet": "Trecho de exemplo da mensagem",
      "body_text": "Corpo de exemplo da mensagem.\n\nAtenciosamente,\nCliente Exemplo",
      "from": {
        "address": "cliente@example.net",
        "name": "Cliente Exemplo",
        "internal": false
      },
      "to": [
        {
          "address": "vendedora@example.com",
          "name": "Vendedora Exemplo",
          "internal": true
        }
      ],
      "cc": [
        {
          "address": "financeiro@example.net",
          "name": null,
          "internal": false
        }
      ],
      "attachments": [
        {
          "filename": "exemplo.pdf",
          "mime_type": "application/pdf",
          "size_bytes": 182400
        }
      ]
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": false,
    "next_cursor": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped email message page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List email conversations

Source: https://docs.salesbud.com.br/reference/emails/listemails

get <code>{"/v1/emails"}</code>

Conversations are deduplicated across every mailbox connected to the company: two sellers on the same thread yield one conversation with two mailboxes. Ordered by `last_message_at` ascending (oldest activity first); pages forward with a signed keyset cursor. Message bodies live under `/v1/emails/{email_id}/messages` and need `emails.content.read`.

### Authorization

Requires the `emails.read` scope.

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |
| `last_message_after` | `string date-time` | No | Return conversations whose latest message is after this instant. Filters on activity, not on record changes: a conversation that gains a message moves forward and will be seen again. |
| `last_message_before` | `string date-time` | No | Return conversations whose latest message is before this instant. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/emails?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "eml_01JEXAMPLE",
      "object": "email_conversation",
      "subject": "Assunto de exemplo",
      "first_message_at": "2026-01-10T12:00:00.000Z",
      "last_message_at": "2026-01-12T15:30:00.000Z",
      "message_count": 3,
      "has_attachments": true,
      "last_message_direction": "inbound",
      "participants": [
        {
          "address": "vendedora@example.com",
          "name": "Vendedora Exemplo",
          "internal": true
        },
        {
          "address": "cliente@example.net",
          "name": "Cliente Exemplo",
          "internal": false
        }
      ],
      "mailboxes": [
        {
          "address": "vendedora@example.com",
          "owner": {
            "id": "usr_01JEXAMPLE",
            "name": "Vendedora Exemplo",
            "email": "vendedora@example.com"
          }
        },
        {
          "address": "vendas@example.com",
          "owner": null
        }
      ],
      "accounts": [
        {
          "id": "acc_01JEXAMPLE",
          "domain": "example.net",
          "cnpj": "00.000.000/0001-00"
        }
      ]
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor03"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped email conversation page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get one completed meeting

Source: https://docs.salesbud.com.br/reference/meetings/getmeeting

get <code>{"/v1/meetings/{meeting_id}"}</code>

Returns a single [meeting object](/objects/meeting/).

### Authorization

Requires the `meetings.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": "mtg_ExampleMeeting01",
    "object": "meeting",
    "title": "Discovery - Acme",
    "status": "completed",
    "type": "video",
    "audience": "external",
    "meeting_at": "2026-01-14T13:00:00.000Z",
    "duration_seconds": 2840,
    "no_show": false,
    "participants": [
      {
        "display_name": "Example Buyer",
        "email": "buyer@example.com",
        "kind": "external"
      },
      {
        "email": "procurement@example.com",
        "kind": "external"
      }
    ],
    "owner": {
      "id": "usr_ExampleOwner01",
      "name": "Example Rep",
      "email": "rep@example.com",
      "team": {
        "id": "team_ExampleTeam01",
        "name": "Inside Sales",
        "role": "participant",
        "is_default": true
      }
    },
    "template": {
      "id": "tpl_ExampleTemplate01",
      "name": "Discovery script",
      "answers": [
        {
          "field": "qst_ExampleQuestion01",
          "question": "What pains did the customer describe?",
          "answer": "Manual CRM entry after every call, about two hours a day."
        }
      ]
    },
    "tags": [
      {
        "id": "tag_ExampleTag01",
        "name": "Enterprise"
      }
    ],
    "customer_questions": [
      {
        "question": "Does it integrate with our CRM?",
        "category": "integrations"
      }
    ],
    "competitors": [
      {
        "id": "ctr_ExampleCompetitor01",
        "name": "Example CRM"
      }
    ],
    "enablement": {
      "next_steps": "Send the proposal by Friday covering 5,400 seats.",
      "meeting_type": "qualification",
      "speaking_duration": {
        "users": [
          {
            "user": {
              "id": "usr_ExampleOwner01",
              "name": "Example Rep",
              "email": "rep@example.com"
            },
            "duration_seconds": 1760
          }
        ],
        "others_seconds": 1080
      }
    },
    "accounts": [
      {
        "id": "acc_ExampleAccount01",
        "domain": "acme.example.com",
        "cnpj": "12345678000199"
      }
    ],
    "overall_evaluation": {
      "status": "completed",
      "justification": "Strong discovery, weak on next steps."
    },
    "skill_scores": [],
    "bot_history": [],
    "transcript": {
      "available": true,
      "status": "completed"
    },
    "created_at": "2026-01-14T12:58:47.000Z",
    "updated_at": "2026-01-14T14:10:43.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped meeting. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List the default template answers for a meeting

Source: https://docs.salesbud.com.br/reference/meetings/getmeetinganswers

get <code>{"/v1/meetings/{meeting_id}/answers"}</code>

Returns a list of [answer objects](/objects/answer/).

### Authorization

Requires the `meetings.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/answers" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "ans_ExampleAnswer01",
      "object": "meeting_answer",
      "question": {
        "id": "qst_ExampleQuestion01",
        "text": "What pains did the customer describe?",
        "type": "text",
        "order": 1
      },
      "answer": "Manual CRM entry after every call, about two hours a day.",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Answers generated for the meeting owner context. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get a meeting transcript

Source: https://docs.salesbud.com.br/reference/meetings/getmeetingtranscript

get <code>{"/v1/meetings/{meeting_id}/transcript"}</code>

Returns the enhanced transcript when available, otherwise the original transcript. An unavailable transcript is represented by a successful response with available set to false and its current processing status.

Returns a single [transcript object](/objects/transcript/).

### Authorization

Requires the `meetings.read` and `transcriptions.read` scopes together.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/transcript" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "object": "transcript",
    "meeting_id": "mtg_ExampleMeeting01",
    "available": true,
    "status": "completed",
    "variant": "enhanced",
    "utterances": [
      {
        "speaker": "Example Rep",
        "text": "Walk me through how the team logs a call today.",
        "start_ms": 12400,
        "end_ms": 16800
      },
      {
        "speaker": "buyer@example.com",
        "text": "Manually, right after. It costs us about two hours a day.",
        "start_ms": 17200,
        "end_ms": 23900
      }
    ],
    "created_at": "2026-01-14T14:02:00.000Z",
    "updated_at": "2026-01-14T14:09:12.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped meeting transcript. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get the overall meeting evaluation

Source: https://docs.salesbud.com.br/reference/meetings/getoverallmeetingevaluation

get <code>{"/v1/meetings/{meeting_id}/evaluations/overall"}</code>

Returns a single [evaluation object](/objects/evaluation/).

### Authorization

Requires the `meetings.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/evaluations/overall" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": "evl_ExampleEvaluation01",
    "object": "overall_meeting_evaluation",
    "meeting_id": "mtg_ExampleMeeting01",
    "status": "completed",
    "score": 8,
    "justification": "Strong discovery, weak on next steps.",
    "created_at": "2026-01-14T14:05:00.000Z",
    "updated_at": "2026-01-14T14:10:43.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Overall evaluation state and result. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List completed meetings

Source: https://docs.salesbud.com.br/reference/meetings/listmeetings

get <code>{"/v1/meetings"}</code>

Returns a list of [meeting objects](/objects/meeting/).

### Authorization

Requires the `meetings.read` scope.

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |
| `created_after` | `string date-time` | No | Return records created after this instant. |
| `created_before` | `string date-time` | No | Return records created before this instant. |
| `meeting_after` | `string date-time` | No | Return records whose event instant is after this value. |
| `meeting_before` | `string date-time` | No | Return records whose event instant is before this value. |
| `owner_email` | `string email` | No | Exact case-insensitive owner email match. |
| `type` | `video` · `audio` | No | Media the record was captured as. Independent of the resource kind: a call can be video and a meeting can be audio. |
| `audience` | `internal` · `external` | No | Whether the record had participants from outside the company (external) or only internal ones. |
| `has_transcript` | `boolean` | No | Filter by existence of a transcript resource. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/meetings?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "mtg_ExampleMeeting01",
      "object": "meeting",
      "title": "Discovery - Acme",
      "status": "completed",
      "type": "video",
      "audience": "external",
      "meeting_at": "2026-01-14T13:00:00.000Z",
      "duration_seconds": 2840,
      "no_show": false,
      "participants": [
        {
          "display_name": "Example Buyer",
          "email": "buyer@example.com",
          "kind": "external"
        },
        {
          "email": "procurement@example.com",
          "kind": "external"
        }
      ],
      "owner": {
        "id": "usr_ExampleOwner01",
        "name": "Example Rep",
        "email": "rep@example.com",
        "team": {
          "id": "team_ExampleTeam01",
          "name": "Inside Sales",
          "role": "participant",
          "is_default": true
        }
      },
      "template": {
        "id": "tpl_ExampleTemplate01",
        "name": "Discovery script",
        "answers": [
          {
            "field": "qst_ExampleQuestion01",
            "question": "What pains did the customer describe?",
            "answer": "Manual CRM entry after every call, about two hours a day."
          }
        ]
      },
      "tags": [
        {
          "id": "tag_ExampleTag01",
          "name": "Enterprise"
        }
      ],
      "customer_questions": [
        {
          "question": "Does it integrate with our CRM?",
          "category": "integrations"
        }
      ],
      "competitors": [
        {
          "id": "ctr_ExampleCompetitor01",
          "name": "Example CRM"
        }
      ],
      "enablement": {
        "next_steps": "Send the proposal by Friday covering 5,400 seats.",
        "meeting_type": "qualification",
        "speaking_duration": {
          "users": [
            {
              "user": {
                "id": "usr_ExampleOwner01",
                "name": "Example Rep",
                "email": "rep@example.com"
              },
              "duration_seconds": 1760
            }
          ],
          "others_seconds": 1080
        }
      },
      "accounts": [
        {
          "id": "acc_ExampleAccount01",
          "domain": "acme.example.com",
          "cnpj": "12345678000199"
        }
      ],
      "overall_evaluation": {
        "status": "completed",
        "justification": "Strong discovery, weak on next steps."
      },
      "skill_scores": [],
      "bot_history": [],
      "transcript": {
        "available": true,
        "status": "completed"
      },
      "created_at": "2026-01-14T12:58:47.000Z",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor01"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped meeting page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Get a WhatsApp conversation

Source: https://docs.salesbud.com.br/reference/whatsapp/getwhatsappconversation

get <code>{"/v1/whatsapp/{whatsapp_id}"}</code>

Conversation metadata only: contact, seller, linked accounts and activity instants. A `wa_` id issued to another company, or any id with a different prefix, is a 404, not a hint. A deleted chat keeps answering, with `deleted_at` and `deleted_reason` set.

### Authorization

Requires the `whatsapp.read` scope.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `whatsapp_id` | `string` | Yes | — |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/whatsapp/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": {
    "id": "wa_01JEXAMPLE",
    "object": "whatsapp_conversation",
    "source": "whatsapp_web",
    "is_group": false,
    "contact": {
      "phone": "+5511999990001",
      "name": "Cliente Exemplo"
    },
    "seller": {
      "user": {
        "id": "usr_01JEXAMPLE",
        "name": "Vendedora Exemplo",
        "email": "vendedora@example.com"
      },
      "phone": "+5511988880001"
    },
    "accounts": [
      {
        "id": "acc_01JEXAMPLE",
        "domain": "example.net",
        "cnpj": "00.000.000/0001-00"
      }
    ],
    "first_message_at": "2026-09-10T12:00:00.000Z",
    "last_message_at": "2026-09-12T15:30:00.000Z",
    "last_message_direction": "inbound",
    "deleted_at": null,
    "deleted_reason": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped WhatsApp conversation. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List WhatsApp conversations

Source: https://docs.salesbud.com.br/reference/whatsapp/listwhatsappconversations

get <code>{"/v1/whatsapp"}</code>

One conversation per chat of a company seller, from WhatsApp Web or RD Conversas. Ordered by `last_message_at` ascending (oldest activity first); pages forward with a signed keyset cursor. `last_message_at` moves only when a message arrives: an edit or a deletion does not move the conversation, so re-read messages to observe them. Message content lives under `/v1/whatsapp/{whatsapp_id}/messages` and needs `whatsapp.content.read`.

### Authorization

Requires the `whatsapp.read` scope.

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |
| `last_message_after` | `string date-time` | No | Return conversations whose latest message is after this instant. Filters on activity, not on record changes: a conversation that gains a message moves forward and will be seen again. |
| `last_message_before` | `string date-time` | No | Return conversations whose latest message is before this instant. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/whatsapp?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "wa_01JEXAMPLE",
      "object": "whatsapp_conversation",
      "source": "whatsapp_web",
      "is_group": false,
      "contact": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "seller": {
        "user": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        },
        "phone": "+5511988880001"
      },
      "accounts": [
        {
          "id": "acc_01JEXAMPLE",
          "domain": "example.net",
          "cnpj": "00.000.000/0001-00"
        }
      ],
      "first_message_at": "2026-09-10T12:00:00.000Z",
      "last_message_at": "2026-09-12T15:30:00.000Z",
      "last_message_direction": "inbound",
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wa_01JGROUP",
      "object": "whatsapp_conversation",
      "source": "whatsapp_web",
      "is_group": true,
      "contact": {
        "phone": null,
        "name": "Grupo Vendas Exemplo"
      },
      "seller": {
        "user": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        },
        "phone": "+5511988880001"
      },
      "accounts": [],
      "first_message_at": "2026-09-11T09:00:00.000Z",
      "last_message_at": "2026-09-12T18:00:00.000Z",
      "last_message_direction": "outbound",
      "deleted_at": null,
      "deleted_reason": null
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor04"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped WhatsApp conversation page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# List the messages of a WhatsApp conversation

Source: https://docs.salesbud.com.br/reference/whatsapp/listwhatsappmessages

get <code>{"/v1/whatsapp/{whatsapp_id}/messages"}</code>

Messages in the order they were sent, oldest first, with sanitized text, media metadata (no media content) and the audio transcription when one exists. A message deleted for everyone stays in place as a tombstone with `deleted_at` set and every content field null; an edited message carries its current text and `edited_at`. Governed by the `whatsapp_content` rate-limit policy in addition to the client and company policies. The cursor is a keyset over `(sent_at, id)` bound to the conversation.

### Authorization

Requires the `whatsapp.read` and `whatsapp.content.read` scopes together.

### Path parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `whatsapp_id` | `string` | Yes | — |

### Query parameters

| Name | Type | Required | Description |
| --- | --- | --- | --- |
| `limit` | `integer` | No | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | No | Signed cursor returned by the previous page. |

### Example request

```bash
curl "https://api.salesbud.com.br/v1/whatsapp/mtg_01JEXAMPLE/messages?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Example response

```json title="200 OK"
{
  "data": [
    {
      "id": "wamsg_01JEXAMPLE1",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:30:00.000Z",
      "direction": "inbound",
      "author": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "type": "text",
      "text": "Pode me mandar a proposta?",
      "media": null,
      "transcription": null,
      "edited_at": null,
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wamsg_01JEXAMPLE2",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:32:00.000Z",
      "direction": "outbound",
      "author": {
        "phone": "+5511988880001",
        "name": "Vendedora Exemplo"
      },
      "type": "audio",
      "text": null,
      "media": {
        "file_name": null,
        "mime_type": "audio/ogg; codecs=opus",
        "size_bytes": 48213
      },
      "transcription": "Claro, envio ainda hoje por e-mail.",
      "edited_at": null,
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wamsg_01JEXAMPLE3",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:35:00.000Z",
      "direction": "inbound",
      "author": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "type": "text",
      "text": null,
      "media": null,
      "transcription": null,
      "edited_at": null,
      "deleted_at": "2026-09-12T15:36:00.000Z",
      "deleted_reason": "revoked_for_everyone"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": false,
    "next_cursor": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Responses

| Status | Description |
| --- | --- |
| `200` | Company-scoped WhatsApp message page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
This page is generated from the OpenAPI specification. Do not edit it by hand.
:::

---

# Changelog

Source: https://docs.salesbud.com.br/resources/changelog

Breaking changes to a released version are announced before they ship. Additive
changes — a new field, a new endpoint, a new error code — can arrive at any time,
so parse defensively: ignore fields you do not recognise.

## Unreleased — first delivery

Version `v1` has not been released to partners yet. Everything below is the shape
of the first delivery.

**Tooling**

- An official MCP server ships alongside v1: `@salesbud/mcp` on npm, one tool per
  public read operation, for Claude Desktop, Claude Code and Cursor. It is a
  client of this API and adds no capability of its own. See
  [MCP server](/guides/mcp/).

**Resources**

- Calls are their own collection at `/v1/calls`, with `call_` identifiers. A
  record captured from a VoIP integration has `object: "call"`; everything else
  is a `meeting`. Ids do not resolve across collections.
- All five routes are mirrored for both collections: list, retrieve, transcript,
  template answers and overall evaluation.
- A call does not carry `bot_history` at all: a VoIP capture has no recording bot, so the field
  is omitted rather than returned empty, and a schema that validates a call rejects it.
  `enablement.meeting_type` stays and is often `null` — that null is legitimate, unlike the bot.
- WhatsApp conversations of the company's sellers are a collection at `/v1/whatsapp`,
  with `wa_` identifiers and `wamsg_` messages, behind `whatsapp.read` (metadata) and
  `whatsapp.content.read` (sanitized text, media metadata, audio transcriptions).
  `last_message_at` moves only on a new message, so re-read messages to observe edits
  and deletions; deleted messages stay as tombstones. Date filters compare in whole
  seconds. See [WhatsApp](/reference/whatsapp/listwhatsappconversations/).

**Errors**

- The error envelope is `{ type, title, detail, code, request_id }`. Every field is
  snake_case; the HTTP status stays in the status line and is not repeated in the body.
  Branch on `code`.
- `POST /oauth/token` keeps the OAuth 2.0 error shape required by RFC 6749 §5.2.

**Pagination and filters**

- Cursors are signed and ordered by ascending id; they bind the filters of the
  request that produced them.
- `updated_after`, `updated_before` and `snapshot_at` were removed. Without a
  dedicated visibility timestamp there is no honest incremental sync; re-query by
  period instead. `updated_at` remains in responses as informative only.
- The date filters are `meeting_after` and `meeting_before`. Earlier drafts also accepted
  `meeting_at_from` and `meeting_at_to`; those never reached a partner and were removed rather
  than shipped as deprecated on day one.

**Access**

- The API is gated by the `API_ACCESS` company feature, checked at token issuance
  and at credential creation.
- Emails and account domains are normalised to lower case in every field, so an
  address can be used as a join key regardless of where it appears.

---

# For LLMs and agents

Source: https://docs.salesbud.com.br/resources/for-agents

This documentation is published in machine-readable form as well as HTML, so an
agent can consume it without scraping markup.

## Entry points

| File | Contents |
| --- | --- |
| [`/llms.txt`](/llms.txt) | Index: one line per page, with its URL and a one-line summary. Start here. |
| [`/llms-full.txt`](/llms-full.txt) | Every page concatenated as plain Markdown, in reading order. |
| [`/openapi.yaml`](/openapi.yaml) | The normative contract. Everything in the API reference is generated from it. |

The format follows the [llms.txt convention](https://llmstxt.org/).

## Any page as Markdown

Append `.md` to any documentation URL to get its source instead of the rendered
page:

```
https://docs.salesbud.com.br/guides/pagination/    → HTML
https://docs.salesbud.com.br/guides/pagination.md  → Markdown
```

Both language versions are available; the pt-BR pages live under `/pt-br/`.

## If you only need to read data from an agent

This API serves an MCP server at `https://api.salesbud.com.br/mcp`, which claude.ai reaches
as a custom connector. It exposes every public read operation as a tool, and
handles token renewal, retries and transcript windowing for you. That is the
shortest path from a credential to an agent reading meetings — see
[MCP server](/guides/mcp/).

Write an HTTP client instead when you are building a service rather than driving
a model.

## If you are generating integration code

Three things are easy to get wrong and worth reading before you write anything:

1. **Loop on `pagination.has_more`, never on `data.length`.** A page can be short
   or empty while more data remains. See [Pagination](/guides/pagination/).
2. **Branch on `error.code`, never on `error.detail`.** The detail is human text
   and may be reworded. See [Errors](/guides/errors/).
3. **Do not infer the resource kind from `type`.** `object` says whether it is a
   `meeting` or a `call`; `type` only says `video` or `audio`, and the two are
   independent. See [Meetings and calls](/guides/meetings-and-calls/).

There is no refresh token — renewal is calling `/oauth/token` again. That is a
property of the client-credentials grant, not an omission.

---

# API Salesbud

Source: https://docs.salesbud.com.br/pt-br

Reunião gravada pelo bot e ligação capturada por VoIP são recursos
    diferentes, então vivem em caminhos diferentes: `/v1/meetings` e
    `/v1/calls`. Mesma forma, identificadores próprios.
  
  
    A credencial pertence a uma empresa, nunca a um usuário. Toda leitura é
    filtrada por ela — não existe parâmetro que amplie esse escopo.
  
  
    Recursos são endereçados por ids opacos com prefixo (`mtg_`, `call_`,
    `usr_`). Ids internos sequenciais nunca são expostos.
  
  
    Toda página existe em Markdown puro, além de [`llms.txt`](/llms.txt) e
    [`llms-full.txt`](/llms-full.txt) para ferramentas que leem documentação
    diretamente.
  

## Sua primeira requisição

Troque a credencial por um token de acesso e leia uma página de reuniões.

```bash title="Primeira requisição"
# 1. Obtenha um token de acesso (vale 1 hora)
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d "client_id=$SALESBUD_CLIENT_ID" \
  -d "client_secret=$SALESBUD_CLIENT_SECRET"

# 2. Leia as reuniões concluídas de um período
curl "https://api.salesbud.com.br/v1/meetings?meeting_after=2026-01-01T00:00:00Z&limit=50" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

A resposta é uma página de reuniões mais um cursor:

```json title="200 OK"
{
  "data": [
    {
      "id": "mtg_H9_chV2YG6UGE0n31AqvDQ",
      "object": "meeting",
      "title": "Discovery — Acme",
      "status": "completed",
      "type": "video",
      "audience": "external",
      "meeting_at": "2026-01-14T13:00:00.000Z",
      "duration_seconds": 2840,
      "participants": [{ "email": "comprador@acme.com", "kind": "external" }],
      "owner": { "id": "usr_CKAyHd20jJb6GGJqhNF9vA", "email": "vendedor@suaempresa.com" },
      "transcript": { "available": true, "status": "completed" }
    }
  ],
  "pagination": { "limit": 50, "has_more": true, "next_cursor": "cur_..." }
}
```

## Por onde seguir

| Se você quer… | Leia |
| --- | --- |
| Fazer a primeira chamada de ponta a ponta | [Começo rápido](/pt-br/get-started/quickstart/) |
| Entender credenciais, tokens e escopos | [Autenticação](/pt-br/guides/authentication/) |
| Saber quando um registro é `meeting` e quando é `call` | [Reuniões e ligações](/pt-br/guides/meetings-and-calls/) |
| Percorrer todo o histórico com segurança | [Paginação](/pt-br/guides/pagination/) |
| Tratar falhas corretamente | [Erros](/pt-br/guides/errors/) |
| Ver cada endpoint, campo e status | [Referência da API](/pt-br/reference/meetings/listmeetings/) |

:::note[O acesso é liberado por empresa]
A API é gateada pela feature `API_ACCESS` da empresa. Se a sua não tiver a
feature ligada, a emissão de token falha com `403 API_ACCESS_DISABLED` — fale com
seu contato na Salesbud em vez de repetir a chamada.
:::

---

# O objeto resposta

Source: https://docs.salesbud.com.br/pt-br/objects/answer

Devolvido pelas rotas de respostas de template, um item por pergunta do template padrão do registro.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque answer identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>meeting_answer</code>
  </div>

Resource type. Always meeting_answer, including for calls.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>object</code>
  </div>

The question this answers.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.id</code>
    <code>string</code>
  </div>

Opaque question identifier, stable across records that share the template. It is the field to key on, not the text, which the company can reword.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.text</code>
    <code>string</code>
  </div>

Question text as the template asks it, sanitized.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.type</code>
    <code>text</code> · <code>option</code> · <code>multi_select</code> · <code>boolean</code> · <code>number</code>
  </div>

What the template expects as an answer. The answer itself always arrives as a string, so this is what says how to parse it.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question.order</code>
    <code>integer</code>
  </div>

Position of the question in the template, starting at zero.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answer</code>
    <code>string</code>
  </div>

Generated answer. A question that produced none is left out of the list rather than returned with a null answer.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
  </div>

When the answer last changed.

</div>

</div>

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# O objeto ligação

Source: https://docs.salesbud.com.br/pt-br/objects/call

Devolvido por todas as rotas de `/v1/calls`. A ligação responde às mesmas cinco rotas de uma reunião e carrega os mesmos campos, exceto `bot_history`; o resto do que muda é quais registros a coleção guarda e como eles costumam vir.

Todo campo está documentado em [o objeto reunião](/pt-br/objects/meeting/). O que segue é só o que difere.

## O que difere

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
  </div>

**`call`, nunca `meeting`.** É o campo que diz qual recurso você tem em mãos.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
  </div>

**Prefixo `call_`.** Um id `mtg_` não resolve em /v1/calls, e o contrário também é 404.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">participants</code>
  </div>

**Em geral um item só, com `phone`.** O registro de VoIP identifica a outra ponta pelo número, não por e-mail.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">type</code>
  </div>

**Em geral `audio`, mas não sempre.** Existem ligações em vídeo; quem distingue o recurso é `object`, nunca `type`.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">enablement.meeting_type</code>
  </div>

**Costuma vir `null`.** O classificador foi feito para reunião de venda e não rotula ligação.

</div>

</div>

:::note[Campos que a ligação não tem]
`bot_history` não está neste objeto: captura de VoIP não tem bot de gravação, então o campo
nunca se aplica e a resposta o omite em vez de mandar array vazio. Um schema que valida ligação
recusa o campo.

`enablement.meeting_type` fica, e costuma vir `null`. Esse nulo é legítimo — o classificador de
estágio de venda não rotula ligação hoje, mas poderia. Não é o mesmo caso do bot.
:::

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# O objeto avaliação

Source: https://docs.salesbud.com.br/pt-br/objects/evaluation

Devolvido pelas rotas de avaliação geral. Todo campo além de `object` e `status` é nulo até a avaliação concluir.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
    aceita null
  </div>

Opaque evaluation ID, null while no evaluation exists.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>overall_meeting_evaluation</code>
  </div>

Resource type. Always overall_meeting_evaluation, including for calls.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_id</code>
    <code>string</code>
  </div>

The record this evaluation belongs to. Carries the prefix of its collection: mtg_ or call_.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>not_started</code> · <code>processing</code> · <code>completed</code> · <code>failed</code>
  </div>

Public processing state.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">score</code>
    <code>integer</code>
    aceita null
  </div>

Score from 0 to 10, or null until the evaluation completes.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">justification</code>
    <code>string</code>
    aceita null
  </div>

Why the score was given, or null until the evaluation completes.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
    aceita null
  </div>

When the evaluation was created, or null when none exists yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
    aceita null
  </div>

When the evaluation last changed, or null when none exists yet.

</div>

</div>

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# O objeto reunião

Source: https://docs.salesbud.com.br/pt-br/objects/meeting

Devolvido por todas as rotas de `/v1/meetings`. A ligação carrega os mesmos campos, exceto `bot_history` — veja [o objeto ligação](/pt-br/objects/call/) para o que muda.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Always prefixed `mtg_`. A `call_` id does not address a meeting: the API refuses it rather than resolving it in the other collection.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>meeting</code>
  </div>

Resource type. A record captured from a VoIP integration is a `call`; everything else is a `meeting`. Independent of `type`, which describes the media: a call can be video and a meeting can be audio.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">title</code>
    <code>string</code>
  </div>

Sanitized meeting title.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>completed</code>
  </div>

Only semantically completed meetings are public in V1.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">type</code>
    <code>video</code> · <code>audio</code>
  </div>

Meeting media type.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">audience</code>
    <code>internal</code> · <code>external</code>
  </div>

Meeting audience classification.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_at</code>
    <code>string date-time</code>
  </div>

Meeting instant in UTC.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">duration_seconds</code>
    <code>integer</code>
  </div>

Non-negative meeting duration.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">no_show</code>
    <code>boolean</code>
  </div>

Whether the completed meeting had no attendee.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">participants</code>
    array of <a href="#meeting-participant">object</a>
  </div>

Normalized external participants.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">owner</code>
    <a href="#meeting-owner">object</a>
  </div>

The user the record belongs to. Company scoping follows this user's default company.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">template</code>
    <a href="#template-reference">object</a>
    aceita null
  </div>

Public meeting template reference when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">tags</code>
    array of <a href="#tag-reference">object</a>
  </div>

Public tag references.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">customer_questions</code>
    array of <a href="#meeting-customer-question">object</a>
  </div>

Customer questions detected in the call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">competitors</code>
    array of <a href="#meeting-competitor">object</a>
  </div>

Competitors mentioned in the call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">enablement</code>
    <a href="#meeting-enablement">object</a>
    aceita null
  </div>

Enablement summary when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">accounts</code>
    array of <a href="#account-reference">object</a>
  </div>

Company-scoped accounts linked to the meeting.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">overall_evaluation</code>
    <code>object</code>
  </div>

Overall evaluation readiness and score summary.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill_scores</code>
    array of <a href="#meeting-skill-score">object</a>
  </div>

Per-skill scores; empty when no skill evaluation exists.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">transcript</code>
    <code>object</code>
  </div>

Transcript availability and its consistent processing state.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
  </div>

UTC instant when the meeting entered Salesbud.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
  </div>

UTC instant of the last change to the meeting record itself. Informational only: it does not advance when a dependent resource such as a transcript, tag or answer changes without touching the meeting row, so it is not a reliable incremental-sync marker.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">bot_history</code>
    array of <a href="#meeting-bot-event">object</a>
  </div>

Recording bot events. Present only on a meeting: a VoIP capture has no bot, so a call does not carry this field at all.

</div>

</div>

## Objetos aninhados

### Meeting participant

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">display_name</code>
    <code>string</code>
    opcional
  </div>

Sanitized external participant name when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    opcional
  </div>

Normalized lowercase email when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">phone</code>
    <code>string</code>
    opcional
  </div>

Normalized phone number when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">kind</code>
    <code>external</code>
  </div>

Participant classification in V1.

</div>

</div>

### Meeting owner

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque owner identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public owner display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    aceita null
  </div>

Normalized owner email, or null when legacy data is malformed.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">team</code>
    <a href="#meeting-owner-team">object</a>
    aceita null
  </div>

Owner default team inside the authenticated company.

</div>

</div>

### Meeting owner team

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque team identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Team display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">role</code>
    <code>participant</code> · <code>leadership</code> · <code>observer</code>
  </div>

Role of the owner in the team. `participant` sees only their own meetings; `leadership` sees the team's and is visible to it; `observer` sees the team's without appearing to the other members.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">is_default</code>
    <code>boolean</code>
  </div>

Whether this is the owner's default team in the company. False when the owner has no default team there and the API reported another one instead, which is why the field exists rather than being implied.

</div>

</div>

### Template reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque public template identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answers</code>
    array of <a href="#meeting-template-answer">object</a>
  </div>

Sanitized template questions and meeting answers.

</div>

</div>

### Meeting template answer

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">field</code>
    <code>string</code>
  </div>

Opaque public template question field identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>string</code>
  </div>

Question text as it was asked, sanitized.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">answer</code>
    <code>string</code>
  </div>

Answer generated for this question. A question that produced none is left out of the list rather than returned with a null answer.

</div>

</div>

### Tag reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque public tag identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public display name.

</div>

</div>

### Meeting customer question

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">question</code>
    <code>string</code>
  </div>

Question the customer asked during the conversation.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">category</code>
    <code>string</code>
  </div>

Category the question was classified into.

</div>

</div>

### Meeting competitor

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque competitor identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Competitor name as registered by the company, not as spoken.

</div>

</div>

### Meeting enablement

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">next_steps</code>
    <code>string</code>
    aceita null
  </div>

Next steps agreed in the conversation, or null when none were identified.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_type</code>
    <code>qualification</code> · <code>negotiation</code> · <code>proposal</code> · <code>closing</code> · <code>technical_meeting</code> · <code>questions_clarifications</code>
    aceita null
  </div>

Sales stage the conversation was classified as, or null. Often null on calls: the classifier was built for sales meetings.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration</code>
    <code>object</code>
  </div>

How long each side spoke.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users</code>
    array of <code>object</code>
  </div>

One entry per company user who spoke. A user who did not speak is left out rather than reported with zero.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users[].user</code>
    <a href="#public-user-reference">object</a>
  </div>

A company user as the API exposes them anywhere other than `owner`: the same identifier, without the team.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.users[].duration_seconds</code>
    <code>integer</code>
  </div>

How long this user spoke, in seconds.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaking_duration.others_seconds</code>
    <code>integer</code>
    aceita null
  </div>

How long everyone outside the company spoke, in seconds, or null when the split was not measured. Not broken down per person: the other side is not identified by user.

</div>

</div>

### Public user reference

A company user as the API exposes them anywhere other than `owner`: the same identifier, without the team.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque user identifier, the same one `owner.id` carries.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">name</code>
    <code>string</code>
  </div>

Public user display name.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">email</code>
    <code>string email</code>
    aceita null
  </div>

Normalized lowercase user email, or null when legacy data is malformed.

</div>

</div>

### Account reference

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">id</code>
    <code>string</code>
  </div>

Opaque account identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">domain</code>
    <code>string</code>
  </div>

Account email domain, lowercased.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">cnpj</code>
    <code>string</code>
    aceita null
  </div>

Brazilian company registration number as stored, with or without punctuation, or null when unknown.

</div>

</div>

### Meeting skill score

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">user</code>
    <a href="#public-user-reference">object</a>
  </div>

The user the score refers to.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill</code>
    <code>object</code>
  </div>

The skill that was scored.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill.id</code>
    <code>string</code>
  </div>

Opaque skill identifier.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">skill.name</code>
    <code>string</code>
  </div>

Skill name as the company registered it.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">score</code>
    <code>integer</code>
  </div>

Score from 0 to 10.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">justification</code>
    <code>string</code>
    aceita null
  </div>

Why the score was given, or null when the evaluation produced none.

</div>

</div>

### Meeting bot event

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">code</code>
    <code>string</code>
    aceita null
  </div>

What happened to the recording bot, as a stable code.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">subcode</code>
    <code>string</code>
    aceita null
  </div>

Further detail on the event, or null when the code carries none.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">occurred_at</code>
    <code>string date-time</code>
  </div>

When the event happened.

</div>

</div>

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# O objeto paginação

Source: https://docs.salesbud.com.br/pt-br/objects/pagination

Acompanha toda resposta de listagem. Itere por `has_more`, nunca pelo tamanho de `data` — veja [Paginação](/pt-br/guides/pagination/).

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">limit</code>
    <code>integer</code>
  </div>

Number of resources requested.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">has_more</code>
    <code>boolean</code>
  </div>

Whether another page is available. A page can be short or empty while this is true, so always continue while it is true instead of relying on the size of data.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">next_cursor</code>
    <code>string</code>
    aceita null
  </div>

Signed opaque cursor for the next page.

</div>

</div>

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# O objeto transcrição

Source: https://docs.salesbud.com.br/pt-br/objects/transcript

Devolvido pelas rotas de transcrição. Transcrição indisponível é resposta bem-sucedida com `available: false`, não erro.

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">object</code>
    <code>transcript</code>
  </div>

Resource type. Always `transcript`, including for a call.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">meeting_id</code>
    <code>string</code>
  </div>

The record this transcript belongs to. Carries the prefix of its collection: `mtg_` or `call_`.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">available</code>
    <code>true</code> · <code>false</code>
  </div>

Whether there is a transcript to read. `false` is a successful response carrying the reason in `status`, not an error.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">status</code>
    <code>completed</code> · <code>not_started</code> · <code>processing</code> · <code>failed</code>
  </div>

Processing state. Always `completed` while `available` is true; otherwise why the transcript is not there yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">variant</code>
    <code>enhanced</code> · <code>original</code>
    aceita null
  </div>

Which transcript was returned: `enhanced` when the improved version exists, `original` otherwise. Null while none is available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">utterances</code>
    array of <a href="#meeting-transcript-utterance">object</a>
  </div>

The transcript itself, in chronological order. Always empty while the transcript is unavailable.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">created_at</code>
    <code>string date-time</code>
    aceita null
  </div>

When the transcript was first stored, or null when none exists yet.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">updated_at</code>
    <code>string date-time</code>
    aceita null
  </div>

When the transcript last changed, or null when none exists yet.

</div>

</div>

## Objetos aninhados

### Meeting transcript utterance

<div class="attrs">

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">speaker</code>
    <code>string</code>
    aceita null
  </div>

Sanitized speaker label when available.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">text</code>
    <code>string</code>
  </div>

Sanitized utterance text.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">start_ms</code>
    <code>integer</code>
  </div>

Utterance start offset in milliseconds.

</div>

<div class="attr">
  <div class="attr-head">
    <code class="attr-name">end_ms</code>
    <code>integer</code>
  </div>

Utterance end offset in milliseconds.

</div>

</div>

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Começo rápido

Source: https://docs.salesbud.com.br/pt-br/get-started/quickstart

Este guia vai de uma credencial nova até uma página de reuniões. São duas
requisições.

## Antes de começar

Você precisa de um **client ID** e um **client secret**, emitidos pela Salesbud
para a sua empresa. A credencial não é self-service: peça ao seu contato na
Salesbud, que também confirma se a feature `API_ACCESS` está ligada.

O secret aparece uma única vez, na criação. Guarde num gerenciador de segredos —
não dá para lê-lo depois, só rotacionar.

1. **Obtenha um token de acesso**

   Os tokens são emitidos por OAuth 2.0 client credentials. Envie a credencial
   como campos form-encoded:

   
   
   ```bash
   curl -X POST https://api.salesbud.com.br/oauth/token \
     -d 'grant_type=client_credentials' \
     -d "client_id=$SALESBUD_CLIENT_ID" \
     -d "client_secret=$SALESBUD_CLIENT_SECRET"
   ```
   
   
   ```js
   const response = await fetch("https://api.salesbud.com.br/oauth/token", {
     method: "POST",
     headers: { "Content-Type": "application/x-www-form-urlencoded" },
     body: new URLSearchParams({
       grant_type: "client_credentials",
       client_id: process.env.SALESBUD_CLIENT_ID,
       client_secret: process.env.SALESBUD_CLIENT_SECRET,
     }),
   });

   const { access_token, expires_in } = await response.json();
   ```
   
   
   ```python
   import os, requests

   response = requests.post(
       "https://api.salesbud.com.br/oauth/token",
       data={
           "grant_type": "client_credentials",
           "client_id": os.environ["SALESBUD_CLIENT_ID"],
           "client_secret": os.environ["SALESBUD_CLIENT_SECRET"],
       },
   )
   access_token = response.json()["access_token"]
   ```
   
   

   O endpoint aceita apenas `application/x-www-form-urlencoded`. Mandar JSON
   devolve `415 UNSUPPORTED_MEDIA_TYPE`.

   ```json title="200 OK"
   {
     "access_token": "eyJhbGciOiJSUzI1NiIs...",
     "token_type": "Bearer",
     "expires_in": 3600,
     "scope": "meetings.read calls.read"
   }
   ```

2. **Confirme o que o token pode fazer**

   `GET /v1/context` informa a empresa, os escopos e o limite de requisições em
   vigor para aquele token. É o jeito mais rápido de separar problema de
   configuração de bug de integração.

   ```bash
   curl https://api.salesbud.com.br/v1/context \
     -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
   ```

3. **Leia uma página de reuniões**

   ```bash
   curl "https://api.salesbud.com.br/v1/meetings?meeting_after=2026-01-01T00:00:00Z&limit=50" \
     -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
   ```

   Só voltam reuniões da sua empresa que terminaram o processamento. Veja
   [Reuniões e ligações](/pt-br/guides/meetings-and-calls/) para entender o que
   é "terminou" e por que os registros de VoIP ficam em `/v1/calls`.

4. **Percorra o resto do histórico**

   Devolva o `next_cursor` sem alterar, mantendo todos os filtros idênticos, e
   itere pelo `has_more` — nunca pelo tamanho de `data`.

   ```js
   let cursor = null;
   do {
     const url = new URL("https://api.salesbud.com.br/v1/meetings");
     url.searchParams.set("meeting_after", "2026-01-01T00:00:00Z");
     url.searchParams.set("limit", "50");
     if (cursor) url.searchParams.set("cursor", cursor);

     const page = await fetch(url, {
       headers: { Authorization: `Bearer ${accessToken}` },
     }).then((r) => r.json());

     for (const meeting of page.data) await handle(meeting);
     cursor = page.pagination.next_cursor;
   } while (cursor);
   ```

   Uma página pode voltar curta — até vazia — com `has_more` em `true`. Isso é
   esperado: o serviço limita quanto varre por requisição. A
   [Paginação](/pt-br/guides/pagination/) explica as garantias.

## Renovando o token

Não existe refresh token. Client credentials é um fluxo máquina-a-máquina, sem
usuário para reconsentir, então a
[RFC 6749 §4.4.3](https://www.rfc-editor.org/rfc/rfc6749#section-4.4.3) diz para
não emitir um. Quando o token expira, chame `/oauth/token` de novo com a mesma
credencial.

Na prática: guarde o token por um pouco menos que o `expires_in` e reemita
também em `401 INVALID_ACCESS_TOKEN`, não só por relógio — uma credencial pode
ser revogada antes de o prazo acabar.

## Se quem lê é um modelo

Se o consumidor é um agente e não um serviço, dá para pular o cliente HTTP: o
[servidor MCP](/pt-br/guides/mcp/) expõe estas mesmas operações como tools para
Claude Desktop, Claude Code e Cursor, e cuida da renovação de token e dos
retries. A credencial acima continua sendo necessária.

## Falhas comuns na primeira chamada

| Resposta | O que significa |
| --- | --- |
| `415 UNSUPPORTED_MEDIA_TYPE` | O pedido de token foi enviado como JSON. Use form encoding. |
| `401 INVALID_CLIENT` | Client ID ou secret errado, ou credencial revogada. |
| `403 API_ACCESS_DISABLED` | A empresa não tem a feature `API_ACCESS`. Repetir não resolve. |
| `403 INSUFFICIENT_SCOPE` | Falta um escopo que a rota exige — `calls.read` para ligações, e `transcriptions.read` somado ao escopo da coleção para transcrição. |
| `404 RESOURCE_NOT_FOUND` | O id não existe, é de outra empresa, ou é da outra coleção. |

---

# Autenticação

Source: https://docs.salesbud.com.br/pt-br/guides/authentication

Toda requisição carrega um bearer token, obtido trocando uma credencial em
`POST /oauth/token`.

## A credencial é uma empresa, não um usuário

A credencial identifica **a sua empresa**. Ela não está ligada a uma pessoa, um
login, um time ou um workspace selecionado, e não herda a permissão de ninguém.
Tudo que o token lê tem o escopo da empresa para a qual foi emitido, e nenhum
parâmetro amplia isso.

É por isso que não existe fluxo de login de usuário: não há usuário.

## Obtendo um token

```bash
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d "client_id=$SALESBUD_CLIENT_ID" \
  -d "client_secret=$SALESBUD_CLIENT_SECRET"
```

O endpoint aceita apenas `application/x-www-form-urlencoded`. A credencial também
pode ir por HTTP Basic em vez de no corpo.

```json title="200 OK"
{
  "access_token": "eyJhbGciOiJSUzI1NiIs...",
  "token_type": "Bearer",
  "expires_in": 3600,
  "scope": "meetings.read calls.read transcriptions.read"
}
```

Os tokens são assinados em RS256 e carregam a empresa, os escopos concedidos e a
versão da credencial.

## Escopos

| Escopo | Permite |
| --- | --- |
| `meetings.read` | Listar e obter reuniões, suas respostas de template e sua avaliação geral. |
| `calls.read` | Listar e obter ligações, suas respostas de template e sua avaliação geral. |
| `transcriptions.read` | Ler transcrições, junto com o escopo da coleção a que a transcrição pertence. Pedido à parte porque a transcrição é o dado mais sensível que a API serve. |

A transcrição exige dois escopos: `meetings.read` e `transcriptions.read` para a
transcrição de uma reunião, `calls.read` e `transcriptions.read` para a de uma
ligação. `transcriptions.read` sozinho não lê nada.

O token recebe os escopos anexados à credencial. Chamar uma rota fora deles
devolve `403 INSUFFICIENT_SCOPE` — é o portão funcionando, não rota faltando.

Use `GET /v1/context` para ver os escopos que um token realmente carrega.

## Tempo de vida e renovação

Os tokens expiram depois de `expires_in` segundos (3600 por padrão).

**Não existe refresh token, por desenho.** Client credentials é um fluxo
máquina-a-máquina, sem usuário para reconsentir, então a
[RFC 6749 §4.4.3](https://www.rfc-editor.org/rfc/rfc6749#section-4.4.3) diz para
não emitir um — seria um segundo segredo de vida longa com o mesmo poder do
primeiro. Renovar é chamar `/oauth/token` de novo.

Guarde o token um pouco abaixo do `expires_in` e reemita também quando uma
requisição devolver `401 INVALID_ACCESS_TOKEN`. Uma credencial pode ser revogada
antes de o prazo acabar; só o relógio não percebe.

## Rotacionando um secret

A rotação emite um secret novo e mantém o anterior válido por **sete dias**, para
você fazer deploy do valor novo sem janela de erro.

No máximo dois secrets ficam válidos ao mesmo tempo. Revogue o que está expirando
antes de iniciar outra rotação, senão a segunda é recusada com
`409 ROTATION_IN_PROGRESS`.

Revogar ou desabilitar o client inteiro invalida os tokens emitidos na hora — é
essa a alavanca em caso de suspeita de vazamento, não a rotação.

## Allowlist de IP

A credencial pode ter uma lista opcional de CIDRs IPv4/IPv6. Quando definida, ela
é aplicada tanto na emissão do token quanto em toda requisição que o usa, então
um token vazado não serve fora da sua rede.

:::caution[O secret aparece uma vez]
Ele é guardado só como hash HMAC-SHA256 com pepper do serviço e não pode ser
lido de volta. Se for perdido, rotacione — não há caminho de recuperação.
:::

---

# Erros

Source: https://docs.salesbud.com.br/pt-br/guides/errors

Os erros usam um envelope só, em toda a API:

```json
{
  "error": {
    "type": "ForbiddenError",
    "title": "Forbidden",
    "detail": "The transcriptions.read scope is required.",
    "code": "INSUFFICIENT_SCOPE",
    "request_id": "req_9bfbe80ac1f24e0a9d5b1e5e7a0f2c11"
  }
}
```

## Ramifique pelo `code`

O `code` é o campo estável, legível por máquina. Ele não muda com redação,
tradução ou refatoração.

`title` e `detail` são texto humano e **podem ser reescritos** entre versões. O
`type` é a classe do erro e é estável, mas mais grosseiro que o `code` — vários
códigos compartilham uma classe. Nunca faça parse do `detail`.

Cite o `request_id` ao acionar o suporte: ele identifica a requisição exata no
nosso log.

## Códigos

| HTTP | `code` | Significado |
| --- | --- | --- |
| 400 | `INVALID_LIMIT` | Tamanho de página fora de 1–100. |
| 400 | `INVALID_DATETIME` | Um carimbo de tempo não é RFC 3339 com offset explícito. |
| 400 | `INVALID_FILTER` | Valor ou combinação de filtro não suportada. |
| 400 | `INVALID_CURSOR` | O cursor foi editado, ou os filtros mudaram no meio da travessia. |
| 400 | `INVALID_REQUEST` | Falha de validação, incluindo parâmetro de query desconhecido. |
| 401 | `MISSING_ACCESS_TOKEN` | Nenhum bearer token foi enviado. |
| 401 | `INVALID_ACCESS_TOKEN` | O token é inválido, expirou, ou a credencial foi revogada. |
| 403 | `INSUFFICIENT_SCOPE` | O token não carrega o escopo que a rota exige. |
| 403 | `API_ACCESS_DISABLED` | A empresa não tem acesso à API habilitado. |
| 404 | `RESOURCE_NOT_FOUND` | Não existe, é de outra empresa, ou foi pedido na coleção errada. |
| 413 | `PAYLOAD_TOO_LARGE` | O corpo da requisição excede o tamanho aceito. |
| 415 | `UNSUPPORTED_MEDIA_TYPE` | Content type errado — em geral JSON enviado ao `/oauth/token`. |
| 429 | `RATE_LIMIT_EXCEEDED` | Uma política de rate limit recusou a requisição. |
| 503 | `SERVICE_UNAVAILABLE` | Dependência fail-closed indisponível. Repita com backoff. |
| 503 | `AUDIT_UNAVAILABLE` | Uma leitura bem-sucedida não pôde ser gravada na trilha de auditoria, então não foi devolvida. |

## Identificador desconhecido é sempre 404

Um id de outra empresa devolve `404 RESOURCE_NOT_FOUND`, não `403`. A API não
confirma que o recurso existe em outro lugar — um `403` entregaria esse fato a
quem estivesse sondando ids.

## O endpoint de token é diferente

`POST /oauth/token` devolve o formato de erro do OAuth 2.0 exigido pela
[RFC 6749 §5.2](https://www.rfc-editor.org/rfc/rfc6749#section-5.2), para que
clientes OAuth de mercado continuem funcionando:

```json
{
  "error": "invalid_client",
  "error_description": "Client authentication failed."
}
```

## O que nunca aparece num erro

Stack trace, caminho de arquivo, token, secret, corpo da requisição e payload de
domínio nunca são devolvidos. Uma falha inesperada devolve mensagem genérica fixa
com um `request_id`; o detalhe fica no nosso log.

## Quando repetir

| Status | Repetir? |
| --- | --- |
| `400`, `403`, `404`, `413`, `415` | Não. A requisição vai falhar igual. |
| `401` | Uma vez, depois de reemitir o token. |
| `429` | Sim, depois do `Retry-After`. |
| `503` | Sim, com backoff exponencial e jitter. |

---

# Servidor MCP

Source: https://docs.salesbud.com.br/pt-br/guides/mcp

Esta API serve um servidor [Model Context
Protocol](https://modelcontextprotocol.io) por HTTP em `https://api.salesbud.com.br/mcp`.
Cadastrado uma vez como conector personalizado no claude.ai, ele deixa o modelo
ler reuniões, ligações, transcrições, respostas do template, avaliações, e-mails e
conversas de WhatsApp com a sua credencial — sem nada para instalar.

Ele é um cliente da API pública, não uma superfície nova: mesma credencial
OAuth, mesmos escopos, mesmos limites de uso e **apenas operações de leitura** —
nada que ele alcance é diferente do que a API HTTP devolveria para a mesma
credencial.

## Antes de começar

Você precisa de uma credencial de API com os escopos que pretende usar, da
feature `API_ACCESS` habilitada para a sua empresa e de acesso de administrador
na organização do Claude que vai guardar o conector. As credenciais não são
self-service; veja o [Começo rápido](/pt-br/get-started/quickstart/).

## Cadastrar o conector

Quem administra a organização no Claude cadastra um conector personalizado
apontando para `https://api.salesbud.com.br/mcp`. A autenticação é **No sign-in** — este
servidor não tem fluxo OAuth — e a credencial da empresa cujos dados o conector
deve ler viaja em dois headers de requisição:

| Header            | Valor           |
| ----------------- | --------------- |
| `X-Client-Id`     | `sb_client_...` |
| `X-Client-Secret` | `sb_secret_...` |

Qualquer outro cliente HTTP pode mandar a mesma credencial como Basic (RFC 7617):

```
Authorization: Basic <base64 de client_id:client_secret>
```

Duas consequências de uma credencial cadastrada uma vez pelo administrador:

- **Ela fica guardada do lado da Anthropic e é compartilhada pela organização
  inteira no Claude.** Todo mundo ali lê o que os escopos permitirem, e a trilha
  de auditoria registra a credencial, não quem pediu.
- **Revogar significa rotacionar.** Rotacionar o segredo quebra junto qualquer
  outra integração que use a mesma credencial, então dê uma só para o conector.

Se a credencial restringe acesso por IP, a allowlist precisa conter a faixa de
saída da Anthropic, `160.79.104.0/21`: a requisição chega nesta API de lá, não
do navegador do usuário.

## Tools

Uma tool por operação pública de leitura. A coluna de escopo é o que a
credencial precisa carregar — escopo faltando volta como uma mensagem legível de
`INSUFFICIENT_SCOPE`, que o modelo consegue relatar, não como exceção.

| Tool                             | Escopo                                    |
| -------------------------------- | ----------------------------------------- |
| `get_api_context`                | nenhum além de uma credencial válida      |
| `list_meetings`                  | `meetings.read`                           |
| `get_meeting`                    | `meetings.read`                           |
| `get_meeting_overall_evaluation` | `meetings.read`                           |
| `get_meeting_answers`            | `meetings.read`                           |
| `get_meeting_transcript`         | `meetings.read` + `transcriptions.read`   |
| `list_calls`                     | `calls.read`                              |
| `get_call`                       | `calls.read`                              |
| `get_call_overall_evaluation`    | `calls.read`                              |
| `get_call_answers`               | `calls.read`                              |
| `get_call_transcript`            | `calls.read` + `transcriptions.read`      |
| `list_emails`                    | `emails.read`                             |
| `get_email`                      | `emails.read`                             |
| `list_email_messages`            | `emails.read` + `emails.content.read`     |
| `list_whatsapp_conversations`    | `whatsapp.read`                           |
| `get_whatsapp_conversation`      | `whatsapp.read`                           |
| `list_whatsapp_messages`         | `whatsapp.read` + `whatsapp.content.read` |

Peça `get_api_context` primeiro quando não souber o que a credencial alcança:
ele responde com o client, a empresa, os escopos e o limite por minuto.

O `/oauth/token` não é tool de propósito — o servidor emite e renova o token por
dentro, e expor isso gastaria o bucket de OAuth e colocaria o token na conversa.
As sondas de liveness e readiness também não são tools.

## Transcrições vêm em janelas

Uma transcrição de uma hora tem milhares de falas e não cabe no contexto do
modelo de uma vez. As tools de transcrição devolvem 200 falas por chamada e
dizem quantas faltam; o modelo passa `offset` para continuar. A API devolve a
transcrição inteira — o janelamento é do servidor MCP.

Registro sem transcrição não é erro: a tool responde
`transcript unavailable (status: not_started)`, que é o contrato descrito em
[Reuniões e ligações](/pt-br/guides/meetings-and-calls/).

## Corpo de e-mail custa mais que metadado

`list_emails` e `get_email` devolvem metadado da conversa — participantes, caixas
conectadas, contas vinculadas, contagens — e nunca o corpo. Ler as mensagens é outra
tool, `list_email_messages`, e ela pede mais: `emails.content.read` além de
`emails.read`, e uma política de rate limit própria, mais apertada que a do resto da
API. Peça uma página por vez em vez de varrer a caixa inteira.

Corpo com mais de 2.000 caracteres é cortado no texto que o modelo lê e vai inteiro
no `structuredContent`. Conteúdo de anexo, corpo em HTML e `bcc` a API nunca serve,
então nenhuma tool alcança.

## Texto de WhatsApp custa mais que metadado

`list_whatsapp_conversations` e `get_whatsapp_conversation` devolvem metadado da conversa —
contato, vendedor, contas vinculadas, instantes de atividade — e nunca uma mensagem. Ler as
mensagens é outra tool, `list_whatsapp_messages`, e ela pede mais: `whatsapp.content.read`
além de `whatsapp.read`, e uma política de rate limit própria. Peça uma página por vez em vez
de varrer o histórico inteiro.

Texto ou transcrição de áudio com mais de 2.000 caracteres é cortado no texto que o modelo lê
e vai inteiro em `structuredContent`. Mensagem apagada para todos fica como lápide, sem
conteúdo; mensagem editada mostra o texto atual. Conteúdo de mídia, identificadores do
provedor, reações e estado de leitura nunca chegam ao modelo — a API não os serve. Veja a
[referência de WhatsApp](/pt-br/reference/whatsapp/listwhatsappconversations/).

## O que o servidor resolve, e o que não

Ele resolve renovação de token com margem de segurança, uma repetição em `401`
com token novo, `Retry-After` em `429`, backoff em `503` e falha imediata nos
outros `4xx`. Os erros chegam ao modelo com `code` e `request_id`, então um
filtro de data mal formatado volta como `INVALID_DATETIME` e o modelo corrige o
formato sozinho.

Ele **não** pagina por você. O modelo segue `has_more` e `next_cursor` como
qualquer outro cliente, e a descrição da tool avisa disso, porque
[página curta não é a última página](/pt-br/guides/pagination/).

Também não existe sync incremental, porque a v1 não tem: janele por `meeting_at`
ou `created_at` e reprocesse de forma idempotente por `id`.

<Aside type="caution" title="Dê ao conector uma credencial só dele">
  Uma empresa pode ter 20 clients ativos. Emita uma credencial exclusiva para o
  conector, com apenas os escopos de que ele precisa, para que revogá-la na tela
  de credenciais da API nunca mexa nas suas outras integrações.
</Aside>

## Cota compartilhada

Todas as credenciais de uma empresa dividem o mesmo bucket de 600 requisições
por minuto. Um agente percorrendo um histórico longo consome a cota que as suas
outras integrações estão usando; veja [Limites de uso](/pt-br/guides/rate-limits/).

---

# Reuniões e ligações

Source: https://docs.salesbud.com.br/pt-br/guides/meetings-and-calls

Uma gravação capturada pelo bot de reunião e uma gravação capturada por
integração de VoIP são coisas diferentes no produto, então são recursos
diferentes na API.

| | Reuniões | Ligações |
| --- | --- | --- |
| Coleção | `/v1/meetings` | `/v1/calls` |
| `object` | `meeting` | `call` |
| Prefixo do id | `mtg_` | `call_` |
| De onde vem | Bot de reunião, upload na plataforma, app mobile | Integração de VoIP |

As duas têm as mesmas cinco rotas — listar, obter, transcrição, respostas de
template e avaliação geral — e carregam os mesmos campos, exceto `bot_history`,
que só a reunião tem. O resto do que muda é a coleção e o prefixo do
identificador.

## Não deduza o tipo pelo tipo de mídia

`object` e `type` são eixos independentes:

- `object` diz **o que o registro é** — `meeting` ou `call`.
- `type` diz **como foi capturado** — `video` ou `audio`.

Uma ligação pode ser em vídeo, e uma reunião pode ser em áudio. Numa empresa em
produção, 3 das 49 ligações são vídeo e 6 das 863 reuniões são áudio. Decidir
pelo `type` erra justamente esses casos.

```json
{ "id": "call_IGmhT13We_ltOPvjNSCHKw", "object": "call", "type": "audio" }
{ "id": "call_J781tlst89klU_mWPMXgDA", "object": "call", "type": "video" }
```

## Identificadores não atravessam coleções

Um id endereça um recurso em exatamente uma coleção. Pedir na outra é `404`, não
redirecionamento:

```bash
curl ".../v1/calls/mtg_H9_chV2YG6UGE0n31AqvDQ" -H "Authorization: Bearer $TOKEN"
# 404 RESOURCE_NOT_FOUND
```

É de propósito: evita que um id resolva silenciosamente para um recurso de outro
tipo, e faz a requisição errada aparecer no seu log em vez de devolver algo
plausível.

## O que fica visível

Um registro aparece na API só quando está **concluído** — quando terminou o
processamento que produz transcrição, respostas e avaliações. Registros gravando,
processando ou com falha não são retornados, em vez de serem expostos com
`status` parcial. Por isso `status` é sempre `completed` na v1.

A posse segue a **empresa padrão do dono** da reunião. Se a empresa padrão de um
usuário muda, o histórico dele acompanha. É um trade-off conhecido de resolver
posse em tempo de leitura; veja [Segurança](/pt-br/guides/security/).

## Escolhendo a coleção

Use `/v1/meetings` para conversas de venda capturadas pelo bot ou enviadas por
upload, e `/v1/calls` para conversas por telefone capturadas via VoIP. Se quiser
as duas, pagine cada coleção separadamente — não existe endpoint combinado,
porque na prática elas têm volumes e retenção diferentes.

:::tip[Campos que se comportam diferente em ligações]
A ligação não carrega `bot_history`: não há bot de gravação numa captura de VoIP, então o campo é
omitido em vez de vir vazio — veja [o objeto ligação](/pt-br/objects/call/).

`participants` numa ligação costuma ter um único item com `phone` em vez de
`email`, porque o registro de VoIP identifica a outra ponta pelo número.
`enablement.meeting_type` costuma vir `null` em ligações: o classificador foi
feito para reunião de venda e não rotula ligação.
:::

---

# Paginação

Source: https://docs.salesbud.com.br/pt-br/guides/pagination

As listagens são paginadas com cursores keyset assinados, ordenados por id
crescente. A página padrão é 50 e o máximo é 100.

```json
{
  "data": [ /* … */ ],
  "pagination": { "limit": 50, "has_more": true, "next_cursor": "cur_..." }
}
```

## Itere pelo `has_more`, nunca por `data.length`

Este é o erro de integração mais comum contra esta API.

Uma página pode voltar **curta, ou até vazia, com `has_more` em `true`**. O
serviço limita quanto varre por requisição, então numa janela esparsa as linhas
restantes vêm na chamada seguinte, em vez de numa resposta única e lenta.

```js
// Certo
let cursor = null;
do {
  const page = await fetchPage(cursor);
  await handle(page.data);
  cursor = page.pagination.next_cursor;
} while (cursor);
```

```js
// Errado — para cedo numa janela esparsa
let page = await fetchPage();
while (page.data.length > 0) { /* … */ }
```

## O cursor amarra seus filtros

O `next_cursor` é assinado e cobre todos os filtros da requisição que o gerou.
Devolva-o **sem alterar**, com os **mesmos filtros**. Mudar um filtro no meio da
travessia, ou editar o cursor, devolve `400 INVALID_CURSOR` em vez de retornar
silenciosamente uma fatia diferente.

Para trocar de filtro, comece uma travessia nova, sem cursor.

## O que a ordenação por id garante

Ordenar por id crescente torna a travessia estável sem snapshot: registros
criados enquanto você pagina recebem ids maiores, então caem depois da sua
posição atual e nunca deslocam uma página que você já leu.

O trade-off é o espelho disso: um registro que fica **concluído** depois de você
já ter passado pelo id dele não entra na travessia em andamento. Ele aparece na
sua próxima passada por aquela janela.

É por isso que a v1 não tem sync incremental — não existe carimbo dedicado de
"ficou visível" que tornasse um delta honesto. O `updated_at` é informativo; não
use como marca d'água de sincronização. Reconsulte por período:

```bash
curl ".../v1/meetings?meeting_after=2026-01-01T00:00:00Z&meeting_before=2026-02-01T00:00:00Z"
```

## Filtros

| Parâmetro | Filtra por |
| --- | --- |
| `meeting_after` / `meeting_before` | Quando a reunião ou ligação aconteceu. É o filtro que a maioria das integrações quer. |
| `created_after` / `created_before` | Quando o registro foi criado na Salesbud. |
| `owner_email` | Dono exato, sem diferenciar maiúsculas. |
| `type` | `video` ou `audio` — a mídia, não o tipo de recurso. |
| `audience` | `internal` ou `external`. |
| `has_transcript` | Se existe recurso de transcrição. |

Todos os carimbos de tempo são RFC 3339 com offset explícito.
`2026-01-01T00:00:00Z` é válido; `2026-01-01` não é, e devolve
`400 INVALID_DATETIME`.

---

# Limites de requisição

Source: https://docs.salesbud.com.br/pt-br/guides/rate-limits

As requisições são medidas por token buckets distribuídos. Vários se aplicam ao
mesmo tempo, e o mais restrito vence.

| Bucket | Aplica-se a | Cota |
| --- | --- | --- |
| `client` | Toda requisição do plano de dados, por credencial | 120/min, ajustável por credencial |
| `company` | Toda requisição do plano de dados, somando as credenciais da empresa | 600/min |
| `transcript` | Só leitura de transcrição, além dos dois acima | 60/min |
| `oauth-client` | `POST /oauth/token`, por client id | 10/min |
| `oauth-ip` | `POST /oauth/token`, por IP de origem | 30/min |

Uma empresa pode ter até 20 credenciais ativas, e juntas elas nunca passam da
cota da empresa. Use `GET /v1/context` para ler a cota de credencial em vigor
para a sua.

## Headers

Toda resposta traz o estado atual. O `RateLimit-Policy` lista todos os buckets
que a requisição consumiu; o `RateLimit` informa apenas o que está mais perto de
esgotar:

```http
RateLimit-Policy: "client";q=120;w=60, "company";q=600;w=60
RateLimit: "client";r=118;t=1
```

`q` é a cota, `w` a janela em segundos, `r` o que resta e `t` os segundos até o
próximo token. O `RateLimit` nomeia o bucket com menos tokens no momento:
normalmente o da sua credencial, o de transcrição nas leituras de transcrição, e
o da empresa quando as outras credenciais estão gastando. Ritme suas chamadas
pelos tetos do `RateLimit-Policy` — o bucket de uma leitura não é o único que te
limita.

O `r` conta tokens no bucket, cuja capacidade é o burst e não a cota, então ele
não começa em `q` e decrementa. Leia como espaço disponível agora.

Uma requisição recusada devolve `429 RATE_LIMIT_EXCEEDED` com `Retry-After` em
segundos. Espere esse tempo — não repita na hora, e adicione jitter para que uma
frota de workers não volte a sincronizar na janela seguinte.

## Comportamento quando o limitador degrada

Os caminhos de leitura **falham abertos**: se o limitador ficar brevemente
inacessível, as leituras de negócio continuam servindo em vez de cair.

Emissão de token, leitura de transcrição e rotas internas de control plane
**falham fechadas** com `503 SERVICE_UNAVAILABLE`. São os caminhos em que servir
tráfego sem medição é pior que não servir.

## O limite do WAF é separado

Há um limite grosseiro por IP no WAF, antes de o tráfego chegar ao serviço, e ele
não faz parte das cotas acima: cerca de 2000 requisições por cinco minutos por IP
de origem, e 100 por cinco minutos no `/oauth/token`.

Ele falha de outro jeito, e isso importa quando você o encontra:

- corta com `403`, não `429`, então parece falha de autorização;
- não vem código de erro, nem `Retry-After`, nem headers de rate limit, porque a
  requisição nunca chega ao serviço;
- conta por IP de origem, então toda integração atrás do mesmo endereço de saída
  divide o mesmo limite, com quais credenciais for.

Um `403` sem nenhum corpo de erro nosso, especialmente no `/oauth/token`, é esse
limite e não falta de escopo.

## Como ficar abaixo do limite

- Pagine com `limit=100` em vez de muitas páginas pequenas — uma requisição
  devolve até 100 registros pelo mesmo custo de uma que devolve 10.
- Filtre por período em vez de percorrer todo o histórico repetidamente.
- Guarde o token de acesso; cada chamada a `/oauth/token` consome bucket próprio.
- Busque transcrição só dos registros que você vai processar. O `has_transcript`
  permite filtrar a listagem antes.

---

# Segurança

Source: https://docs.salesbud.com.br/pt-br/guides/security

## Isolamento por empresa

Toda leitura é filtrada pela empresa que está no token de acesso. A empresa vem
da claim do token, nunca de um parâmetro da requisição, então não existe entrada
capaz de ampliar o escopo de uma credencial.

Os recursos são endereçados por identificadores opacos com prefixo (`mtg_`,
`call_`, `usr_`, `tpl_`), alocados por empresa. Ids internos sequenciais nunca
são expostos, então não dá para inferir volume nem enumerar ids.

Um identificador que não é da sua empresa devolve `404`, nunca `403` — a API não
confirma que o recurso existe em outro lugar.

## A posse é resolvida na leitura

Um registro pertence à **empresa padrão do usuário dono dele**, avaliada no
momento da consulta, e não congelada quando o registro foi criado.

Vale dizer a consequência com todas as letras: **se a empresa padrão de um
usuário muda, o histórico dele vai junto.** Um vendedor que troca de empresa
dentro do mesmo grupo leva o histórico de reuniões para a API da empresa nova.

É um trade-off deliberado de não desnormalizar a posse no registro. Se esse
comportamento não couber no seu acordo de dados, levante isso com seu contato na
Salesbud antes de construir em cima.

## Transcrições têm portão próprio

`transcriptions.read` é um escopo distinto de `meetings.read` e `calls.read`, e
uma credencial pode ter os escopos de coleção sem ele. A transcrição é o dado
mais sensível que a API serve — a fala literal do cliente — então acessar o
metadado de uma reunião não implica acessar o que foi dito nela.

O portão soma em vez de substituir o escopo da coleção: a requisição de
transcrição é checada contra os dois, então uma credencial restrita a ligações
não alcança transcrição de reunião por ter `transcriptions.read`.

O texto da transcrição é sanitizado antes de sair do serviço.

## Tudo é auditado

Toda leitura de negócio é registrada: qual credencial, qual empresa, qual rota,
quais ids de recurso e qual identidade chamou.

A gravação da auditoria é **fail closed**. Se a trilha não puder ser escrita, a
leitura não é devolvida — você recebe `503 AUDIT_UNAVAILABLE`. Uma resposta
bem-sucedida, portanto, significa que o acesso ficou registrado.

## Higiene de credencial

- Os secrets são guardados apenas como hash HMAC-SHA256 com pepper do serviço.
  Não podem ser lidos de volta, só rotacionados.
- A rotação mantém o secret anterior válido por sete dias para você fazer deploy
  sem downtime; revogue assim que o novo estiver no ar.
- Em suspeita de vazamento, **revogue o client** — isso invalida os tokens
  emitidos na hora. A rotação não, porque o secret antigo continua valendo.
- Configure allowlist de IP quando a integração roda de endereços fixos. Ela vale
  na emissão e em toda requisição, então um token vazado não serve de fora.

## O que nunca sai do serviço

Stack trace, caminho de arquivo, id interno, token, secret e corpo de requisição
nunca são devolvidos numa resposta, nem escritos em log.

---

# Issue an access token

Source: https://docs.salesbud.com.br/pt-br/reference/authentication/issueaccesstoken

post <code>{"/oauth/token"}</code>

### Autorização

Sem autenticação.

### Exemplo de requisição

```bash
curl -X POST https://api.salesbud.com.br/oauth/token \
  -d 'grant_type=client_credentials' \
  -d 'client_id=sb_client_...' \
  -d 'client_secret=sb_secret_...'
```

### Exemplo de resposta

```json title="200 OK"
{
  "access_token": "eyJhbGciOiJSUzI1NiJ9.EXAMPLE_PAYLOAD.EXAMPLE_SIGNATURE",
  "token_type": "Bearer",
  "expires_in": 3600,
  "scope": "meetings.read calls.read transcriptions.read"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Access token issued. |
| `400` | OAuth client-credentials request rejected. |
| `401` | OAuth client authentication failed. |
| `413` | The request body is larger than the accepted limit. |
| `415` | The request media type is not supported by this endpoint. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get one completed call

Source: https://docs.salesbud.com.br/pt-br/reference/calls/getcall

get <code>{"/v1/calls/{call_id}"}</code>

Devolve um [objeto ligação](/pt-br/objects/call/).

### Autorização

Exige o escopo `calls.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `call_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": "call_ExampleCall01",
    "object": "call",
    "title": "Call to +55 11 99999-0000",
    "status": "completed",
    "type": "audio",
    "audience": "external",
    "meeting_at": "2026-01-15T18:28:01.000Z",
    "duration_seconds": 714,
    "no_show": false,
    "participants": [
      {
        "phone": "+5511999990000",
        "kind": "external"
      }
    ],
    "owner": {
      "id": "usr_ExampleOwner01",
      "name": "Example Rep",
      "email": "rep@example.com",
      "team": {
        "id": "team_ExampleTeam01",
        "name": "Inside Sales",
        "role": "participant",
        "is_default": true
      }
    },
    "template": {
      "id": "tpl_ExampleTemplate01",
      "name": "Discovery script",
      "answers": [
        {
          "field": "qst_ExampleQuestion01",
          "question": "What pains did the customer describe?",
          "answer": "Wants a demo before renewing the current contract."
        }
      ]
    },
    "tags": [],
    "customer_questions": [],
    "competitors": [],
    "enablement": {
      "next_steps": "Online meeting booked for Thursday at 13:30; send the invite today.",
      "meeting_type": null,
      "speaking_duration": {
        "users": [],
        "others_seconds": 677
      }
    },
    "accounts": [],
    "overall_evaluation": {
      "status": "not_started",
      "justification": null
    },
    "skill_scores": [],
    "transcript": {
      "available": false,
      "status": "not_started"
    },
    "created_at": "2026-01-15T19:22:08.000Z",
    "updated_at": "2026-01-15T19:22:53.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped call. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List the default template answers for a call

Source: https://docs.salesbud.com.br/pt-br/reference/calls/getcallanswers

get <code>{"/v1/calls/{call_id}/answers"}</code>

Devolve uma lista de [objetos resposta](/pt-br/objects/answer/).

### Autorização

Exige o escopo `calls.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `call_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/answers" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "ans_ExampleAnswer01",
      "object": "meeting_answer",
      "question": {
        "id": "qst_ExampleQuestion01",
        "text": "What pains did the customer describe?",
        "type": "text",
        "order": 1
      },
      "answer": "Manual CRM entry after every call, about two hours a day.",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Answers generated for the call owner context. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get a call transcript

Source: https://docs.salesbud.com.br/pt-br/reference/calls/getcalltranscript

get <code>{"/v1/calls/{call_id}/transcript"}</code>

Returns the enhanced transcript when available, otherwise the original transcript. An unavailable transcript is represented by a successful response with available set to false and its current processing status.

Devolve um [objeto transcrição](/pt-br/objects/transcript/).

### Autorização

Exige os escopos `calls.read` e `transcriptions.read` juntos.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `call_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/transcript" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "object": "transcript",
    "meeting_id": "call_ExampleCall01",
    "available": false,
    "status": "processing",
    "variant": null,
    "utterances": [],
    "created_at": null,
    "updated_at": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped call transcript. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get the overall call evaluation

Source: https://docs.salesbud.com.br/pt-br/reference/calls/getoverallcallevaluation

get <code>{"/v1/calls/{call_id}/evaluations/overall"}</code>

Devolve um [objeto avaliação](/pt-br/objects/evaluation/).

### Autorização

Exige o escopo `calls.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `call_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/calls/call_01JEXAMPLE/evaluations/overall" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": null,
    "object": "overall_meeting_evaluation",
    "meeting_id": "call_ExampleCall01",
    "status": "not_started",
    "score": null,
    "justification": null,
    "created_at": null,
    "updated_at": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Overall evaluation state and result. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List completed calls

Source: https://docs.salesbud.com.br/pt-br/reference/calls/listcalls

get <code>{"/v1/calls"}</code>

Calls are the records captured from a VoIP integration, addressed by their own `call_` identifiers. They answer the same five routes as meetings and carry the same fields except `bot_history`: a VoIP capture has no recording bot, so the field is absent, not empty.

Devolve uma lista de [objetos ligação](/pt-br/objects/call/).

### Autorização

Exige o escopo `calls.read`.

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |
| `created_after` | `string date-time` | Não | Return records created after this instant. |
| `created_before` | `string date-time` | Não | Return records created before this instant. |
| `meeting_after` | `string date-time` | Não | Return records whose event instant is after this value. |
| `meeting_before` | `string date-time` | Não | Return records whose event instant is before this value. |
| `owner_email` | `string email` | Não | Exact case-insensitive owner email match. |
| `type` | `video` · `audio` | Não | Media the record was captured as. Independent of the resource kind: a call can be video and a meeting can be audio. |
| `audience` | `internal` · `external` | Não | Whether the record had participants from outside the company (external) or only internal ones. |
| `has_transcript` | `boolean` | Não | Filter by existence of a transcript resource. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/calls?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "call_ExampleCall01",
      "object": "call",
      "title": "Call to +55 11 99999-0000",
      "status": "completed",
      "type": "audio",
      "audience": "external",
      "meeting_at": "2026-01-15T18:28:01.000Z",
      "duration_seconds": 714,
      "no_show": false,
      "participants": [
        {
          "phone": "+5511999990000",
          "kind": "external"
        }
      ],
      "owner": {
        "id": "usr_ExampleOwner01",
        "name": "Example Rep",
        "email": "rep@example.com",
        "team": {
          "id": "team_ExampleTeam01",
          "name": "Inside Sales",
          "role": "participant",
          "is_default": true
        }
      },
      "template": {
        "id": "tpl_ExampleTemplate01",
        "name": "Discovery script",
        "answers": [
          {
            "field": "qst_ExampleQuestion01",
            "question": "What pains did the customer describe?",
            "answer": "Wants a demo before renewing the current contract."
          }
        ]
      },
      "tags": [],
      "customer_questions": [],
      "competitors": [],
      "enablement": {
        "next_steps": "Online meeting booked for Thursday at 13:30; send the invite today.",
        "meeting_type": null,
        "speaking_duration": {
          "users": [],
          "others_seconds": 677
        }
      },
      "accounts": [],
      "overall_evaluation": {
        "status": "not_started",
        "justification": null
      },
      "skill_scores": [],
      "transcript": {
        "available": false,
        "status": "not_started"
      },
      "created_at": "2026-01-15T19:22:08.000Z",
      "updated_at": "2026-01-15T19:22:53.000Z"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor02"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped call page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get the authenticated integration context

Source: https://docs.salesbud.com.br/pt-br/reference/context/getapicontext

get <code>{"/v1/context"}</code>

### Autorização

Sem autenticação.

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/context" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "object": "api_context",
    "client": {
      "id": "sb_client_ExampleClient01",
      "name": "acme-sync"
    },
    "company": {
      "id": "cmp_ExampleCompany01",
      "name": "Example Company"
    },
    "scopes": [
      "meetings.read",
      "calls.read",
      "transcriptions.read"
    ],
    "rate_limit": {
      "requests_per_minute": 120
    }
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | API client, company, scopes and rate-limit context. |
| `401` | Missing or invalid credentials. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get an email conversation

Source: https://docs.salesbud.com.br/pt-br/reference/emails/getemail

get <code>{"/v1/emails/{email_id}"}</code>

Conversation metadata only. An `eml_` id issued to another company, or any id with a different prefix, is a 404, not a hint.

### Autorização

Exige o escopo `emails.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `email_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/emails/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": "eml_01JEXAMPLE",
    "object": "email_conversation",
    "subject": "Assunto de exemplo",
    "first_message_at": "2026-01-10T12:00:00.000Z",
    "last_message_at": "2026-01-12T15:30:00.000Z",
    "message_count": 3,
    "has_attachments": true,
    "last_message_direction": "inbound",
    "participants": [
      {
        "address": "vendedora@example.com",
        "name": "Vendedora Exemplo",
        "internal": true
      },
      {
        "address": "cliente@example.net",
        "name": "Cliente Exemplo",
        "internal": false
      }
    ],
    "mailboxes": [
      {
        "address": "vendedora@example.com",
        "owner": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        }
      },
      {
        "address": "vendas@example.com",
        "owner": null
      }
    ],
    "accounts": [
      {
        "id": "acc_01JEXAMPLE",
        "domain": "example.net",
        "cnpj": "00.000.000/0001-00"
      }
    ]
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped email conversation. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List the messages of an email conversation

Source: https://docs.salesbud.com.br/pt-br/reference/emails/listemailmessages

get <code>{"/v1/emails/{email_id}/messages"}</code>

Messages in the order they were sent, oldest first, with plain-text bodies and attachment metadata (no attachment content). Governed by the `email_content` rate-limit policy in addition to the client and company policies. The cursor is bound to the conversation.

### Autorização

Exige os escopos `emails.read` e `emails.content.read` juntos.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `email_id` | `string` | Sim | — |

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/emails/mtg_01JEXAMPLE/messages?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "emsg_01JEXAMPLE",
      "object": "email_message",
      "conversation_id": "eml_01JEXAMPLE",
      "sent_at": "2026-01-12T15:30:00.000Z",
      "direction": "inbound",
      "subject": "Re: Assunto de exemplo",
      "snippet": "Trecho de exemplo da mensagem",
      "body_text": "Corpo de exemplo da mensagem.\n\nAtenciosamente,\nCliente Exemplo",
      "from": {
        "address": "cliente@example.net",
        "name": "Cliente Exemplo",
        "internal": false
      },
      "to": [
        {
          "address": "vendedora@example.com",
          "name": "Vendedora Exemplo",
          "internal": true
        }
      ],
      "cc": [
        {
          "address": "financeiro@example.net",
          "name": null,
          "internal": false
        }
      ],
      "attachments": [
        {
          "filename": "exemplo.pdf",
          "mime_type": "application/pdf",
          "size_bytes": 182400
        }
      ]
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": false,
    "next_cursor": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped email message page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List email conversations

Source: https://docs.salesbud.com.br/pt-br/reference/emails/listemails

get <code>{"/v1/emails"}</code>

Conversations are deduplicated across every mailbox connected to the company: two sellers on the same thread yield one conversation with two mailboxes. Ordered by `last_message_at` ascending (oldest activity first); pages forward with a signed keyset cursor. Message bodies live under `/v1/emails/{email_id}/messages` and need `emails.content.read`.

### Autorização

Exige o escopo `emails.read`.

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |
| `last_message_after` | `string date-time` | Não | Return conversations whose latest message is after this instant. Filters on activity, not on record changes: a conversation that gains a message moves forward and will be seen again. |
| `last_message_before` | `string date-time` | Não | Return conversations whose latest message is before this instant. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/emails?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "eml_01JEXAMPLE",
      "object": "email_conversation",
      "subject": "Assunto de exemplo",
      "first_message_at": "2026-01-10T12:00:00.000Z",
      "last_message_at": "2026-01-12T15:30:00.000Z",
      "message_count": 3,
      "has_attachments": true,
      "last_message_direction": "inbound",
      "participants": [
        {
          "address": "vendedora@example.com",
          "name": "Vendedora Exemplo",
          "internal": true
        },
        {
          "address": "cliente@example.net",
          "name": "Cliente Exemplo",
          "internal": false
        }
      ],
      "mailboxes": [
        {
          "address": "vendedora@example.com",
          "owner": {
            "id": "usr_01JEXAMPLE",
            "name": "Vendedora Exemplo",
            "email": "vendedora@example.com"
          }
        },
        {
          "address": "vendas@example.com",
          "owner": null
        }
      ],
      "accounts": [
        {
          "id": "acc_01JEXAMPLE",
          "domain": "example.net",
          "cnpj": "00.000.000/0001-00"
        }
      ]
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor03"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped email conversation page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get one completed meeting

Source: https://docs.salesbud.com.br/pt-br/reference/meetings/getmeeting

get <code>{"/v1/meetings/{meeting_id}"}</code>

Devolve um [objeto reunião](/pt-br/objects/meeting/).

### Autorização

Exige o escopo `meetings.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": "mtg_ExampleMeeting01",
    "object": "meeting",
    "title": "Discovery - Acme",
    "status": "completed",
    "type": "video",
    "audience": "external",
    "meeting_at": "2026-01-14T13:00:00.000Z",
    "duration_seconds": 2840,
    "no_show": false,
    "participants": [
      {
        "display_name": "Example Buyer",
        "email": "buyer@example.com",
        "kind": "external"
      },
      {
        "email": "procurement@example.com",
        "kind": "external"
      }
    ],
    "owner": {
      "id": "usr_ExampleOwner01",
      "name": "Example Rep",
      "email": "rep@example.com",
      "team": {
        "id": "team_ExampleTeam01",
        "name": "Inside Sales",
        "role": "participant",
        "is_default": true
      }
    },
    "template": {
      "id": "tpl_ExampleTemplate01",
      "name": "Discovery script",
      "answers": [
        {
          "field": "qst_ExampleQuestion01",
          "question": "What pains did the customer describe?",
          "answer": "Manual CRM entry after every call, about two hours a day."
        }
      ]
    },
    "tags": [
      {
        "id": "tag_ExampleTag01",
        "name": "Enterprise"
      }
    ],
    "customer_questions": [
      {
        "question": "Does it integrate with our CRM?",
        "category": "integrations"
      }
    ],
    "competitors": [
      {
        "id": "ctr_ExampleCompetitor01",
        "name": "Example CRM"
      }
    ],
    "enablement": {
      "next_steps": "Send the proposal by Friday covering 5,400 seats.",
      "meeting_type": "qualification",
      "speaking_duration": {
        "users": [
          {
            "user": {
              "id": "usr_ExampleOwner01",
              "name": "Example Rep",
              "email": "rep@example.com"
            },
            "duration_seconds": 1760
          }
        ],
        "others_seconds": 1080
      }
    },
    "accounts": [
      {
        "id": "acc_ExampleAccount01",
        "domain": "acme.example.com",
        "cnpj": "12345678000199"
      }
    ],
    "overall_evaluation": {
      "status": "completed",
      "justification": "Strong discovery, weak on next steps."
    },
    "skill_scores": [],
    "bot_history": [],
    "transcript": {
      "available": true,
      "status": "completed"
    },
    "created_at": "2026-01-14T12:58:47.000Z",
    "updated_at": "2026-01-14T14:10:43.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped meeting. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List the default template answers for a meeting

Source: https://docs.salesbud.com.br/pt-br/reference/meetings/getmeetinganswers

get <code>{"/v1/meetings/{meeting_id}/answers"}</code>

Devolve uma lista de [objetos resposta](/pt-br/objects/answer/).

### Autorização

Exige o escopo `meetings.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/answers" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "ans_ExampleAnswer01",
      "object": "meeting_answer",
      "question": {
        "id": "qst_ExampleQuestion01",
        "text": "What pains did the customer describe?",
        "type": "text",
        "order": 1
      },
      "answer": "Manual CRM entry after every call, about two hours a day.",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Answers generated for the meeting owner context. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get a meeting transcript

Source: https://docs.salesbud.com.br/pt-br/reference/meetings/getmeetingtranscript

get <code>{"/v1/meetings/{meeting_id}/transcript"}</code>

Returns the enhanced transcript when available, otherwise the original transcript. An unavailable transcript is represented by a successful response with available set to false and its current processing status.

Devolve um [objeto transcrição](/pt-br/objects/transcript/).

### Autorização

Exige os escopos `meetings.read` e `transcriptions.read` juntos.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/transcript" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "object": "transcript",
    "meeting_id": "mtg_ExampleMeeting01",
    "available": true,
    "status": "completed",
    "variant": "enhanced",
    "utterances": [
      {
        "speaker": "Example Rep",
        "text": "Walk me through how the team logs a call today.",
        "start_ms": 12400,
        "end_ms": 16800
      },
      {
        "speaker": "buyer@example.com",
        "text": "Manually, right after. It costs us about two hours a day.",
        "start_ms": 17200,
        "end_ms": 23900
      }
    ],
    "created_at": "2026-01-14T14:02:00.000Z",
    "updated_at": "2026-01-14T14:09:12.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped meeting transcript. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get the overall meeting evaluation

Source: https://docs.salesbud.com.br/pt-br/reference/meetings/getoverallmeetingevaluation

get <code>{"/v1/meetings/{meeting_id}/evaluations/overall"}</code>

Devolve um [objeto avaliação](/pt-br/objects/evaluation/).

### Autorização

Exige o escopo `meetings.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `meeting_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/meetings/mtg_01JEXAMPLE/evaluations/overall" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": "evl_ExampleEvaluation01",
    "object": "overall_meeting_evaluation",
    "meeting_id": "mtg_ExampleMeeting01",
    "status": "completed",
    "score": 8,
    "justification": "Strong discovery, weak on next steps.",
    "created_at": "2026-01-14T14:05:00.000Z",
    "updated_at": "2026-01-14T14:10:43.000Z"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Overall evaluation state and result. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List completed meetings

Source: https://docs.salesbud.com.br/pt-br/reference/meetings/listmeetings

get <code>{"/v1/meetings"}</code>

Devolve uma lista de [objetos reunião](/pt-br/objects/meeting/).

### Autorização

Exige o escopo `meetings.read`.

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |
| `created_after` | `string date-time` | Não | Return records created after this instant. |
| `created_before` | `string date-time` | Não | Return records created before this instant. |
| `meeting_after` | `string date-time` | Não | Return records whose event instant is after this value. |
| `meeting_before` | `string date-time` | Não | Return records whose event instant is before this value. |
| `owner_email` | `string email` | Não | Exact case-insensitive owner email match. |
| `type` | `video` · `audio` | Não | Media the record was captured as. Independent of the resource kind: a call can be video and a meeting can be audio. |
| `audience` | `internal` · `external` | Não | Whether the record had participants from outside the company (external) or only internal ones. |
| `has_transcript` | `boolean` | Não | Filter by existence of a transcript resource. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/meetings?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "mtg_ExampleMeeting01",
      "object": "meeting",
      "title": "Discovery - Acme",
      "status": "completed",
      "type": "video",
      "audience": "external",
      "meeting_at": "2026-01-14T13:00:00.000Z",
      "duration_seconds": 2840,
      "no_show": false,
      "participants": [
        {
          "display_name": "Example Buyer",
          "email": "buyer@example.com",
          "kind": "external"
        },
        {
          "email": "procurement@example.com",
          "kind": "external"
        }
      ],
      "owner": {
        "id": "usr_ExampleOwner01",
        "name": "Example Rep",
        "email": "rep@example.com",
        "team": {
          "id": "team_ExampleTeam01",
          "name": "Inside Sales",
          "role": "participant",
          "is_default": true
        }
      },
      "template": {
        "id": "tpl_ExampleTemplate01",
        "name": "Discovery script",
        "answers": [
          {
            "field": "qst_ExampleQuestion01",
            "question": "What pains did the customer describe?",
            "answer": "Manual CRM entry after every call, about two hours a day."
          }
        ]
      },
      "tags": [
        {
          "id": "tag_ExampleTag01",
          "name": "Enterprise"
        }
      ],
      "customer_questions": [
        {
          "question": "Does it integrate with our CRM?",
          "category": "integrations"
        }
      ],
      "competitors": [
        {
          "id": "ctr_ExampleCompetitor01",
          "name": "Example CRM"
        }
      ],
      "enablement": {
        "next_steps": "Send the proposal by Friday covering 5,400 seats.",
        "meeting_type": "qualification",
        "speaking_duration": {
          "users": [
            {
              "user": {
                "id": "usr_ExampleOwner01",
                "name": "Example Rep",
                "email": "rep@example.com"
              },
              "duration_seconds": 1760
            }
          ],
          "others_seconds": 1080
        }
      },
      "accounts": [
        {
          "id": "acc_ExampleAccount01",
          "domain": "acme.example.com",
          "cnpj": "12345678000199"
        }
      ],
      "overall_evaluation": {
        "status": "completed",
        "justification": "Strong discovery, weak on next steps."
      },
      "skill_scores": [],
      "bot_history": [],
      "transcript": {
        "available": true,
        "status": "completed"
      },
      "created_at": "2026-01-14T12:58:47.000Z",
      "updated_at": "2026-01-14T14:10:43.000Z"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor01"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped meeting page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Get a WhatsApp conversation

Source: https://docs.salesbud.com.br/pt-br/reference/whatsapp/getwhatsappconversation

get <code>{"/v1/whatsapp/{whatsapp_id}"}</code>

Conversation metadata only: contact, seller, linked accounts and activity instants. A `wa_` id issued to another company, or any id with a different prefix, is a 404, not a hint. A deleted chat keeps answering, with `deleted_at` and `deleted_reason` set.

### Autorização

Exige o escopo `whatsapp.read`.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `whatsapp_id` | `string` | Sim | — |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/whatsapp/mtg_01JEXAMPLE" \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": {
    "id": "wa_01JEXAMPLE",
    "object": "whatsapp_conversation",
    "source": "whatsapp_web",
    "is_group": false,
    "contact": {
      "phone": "+5511999990001",
      "name": "Cliente Exemplo"
    },
    "seller": {
      "user": {
        "id": "usr_01JEXAMPLE",
        "name": "Vendedora Exemplo",
        "email": "vendedora@example.com"
      },
      "phone": "+5511988880001"
    },
    "accounts": [
      {
        "id": "acc_01JEXAMPLE",
        "domain": "example.net",
        "cnpj": "00.000.000/0001-00"
      }
    ],
    "first_message_at": "2026-09-10T12:00:00.000Z",
    "last_message_at": "2026-09-12T15:30:00.000Z",
    "last_message_direction": "inbound",
    "deleted_at": null,
    "deleted_reason": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped WhatsApp conversation. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List WhatsApp conversations

Source: https://docs.salesbud.com.br/pt-br/reference/whatsapp/listwhatsappconversations

get <code>{"/v1/whatsapp"}</code>

One conversation per chat of a company seller, from WhatsApp Web or RD Conversas. Ordered by `last_message_at` ascending (oldest activity first); pages forward with a signed keyset cursor. `last_message_at` moves only when a message arrives: an edit or a deletion does not move the conversation, so re-read messages to observe them. Message content lives under `/v1/whatsapp/{whatsapp_id}/messages` and needs `whatsapp.content.read`.

### Autorização

Exige o escopo `whatsapp.read`.

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |
| `last_message_after` | `string date-time` | Não | Return conversations whose latest message is after this instant. Filters on activity, not on record changes: a conversation that gains a message moves forward and will be seen again. |
| `last_message_before` | `string date-time` | Não | Return conversations whose latest message is before this instant. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/whatsapp?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "wa_01JEXAMPLE",
      "object": "whatsapp_conversation",
      "source": "whatsapp_web",
      "is_group": false,
      "contact": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "seller": {
        "user": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        },
        "phone": "+5511988880001"
      },
      "accounts": [
        {
          "id": "acc_01JEXAMPLE",
          "domain": "example.net",
          "cnpj": "00.000.000/0001-00"
        }
      ],
      "first_message_at": "2026-09-10T12:00:00.000Z",
      "last_message_at": "2026-09-12T15:30:00.000Z",
      "last_message_direction": "inbound",
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wa_01JGROUP",
      "object": "whatsapp_conversation",
      "source": "whatsapp_web",
      "is_group": true,
      "contact": {
        "phone": null,
        "name": "Grupo Vendas Exemplo"
      },
      "seller": {
        "user": {
          "id": "usr_01JEXAMPLE",
          "name": "Vendedora Exemplo",
          "email": "vendedora@example.com"
        },
        "phone": "+5511988880001"
      },
      "accounts": [],
      "first_message_at": "2026-09-11T09:00:00.000Z",
      "last_message_at": "2026-09-12T18:00:00.000Z",
      "last_message_direction": "outbound",
      "deleted_at": null,
      "deleted_reason": null
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": true,
    "next_cursor": "cur_ExampleCursor04"
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped WhatsApp conversation page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# List the messages of a WhatsApp conversation

Source: https://docs.salesbud.com.br/pt-br/reference/whatsapp/listwhatsappmessages

get <code>{"/v1/whatsapp/{whatsapp_id}/messages"}</code>

Messages in the order they were sent, oldest first, with sanitized text, media metadata (no media content) and the audio transcription when one exists. A message deleted for everyone stays in place as a tombstone with `deleted_at` set and every content field null; an edited message carries its current text and `edited_at`. Governed by the `whatsapp_content` rate-limit policy in addition to the client and company policies. The cursor is a keyset over `(sent_at, id)` bound to the conversation.

### Autorização

Exige os escopos `whatsapp.read` e `whatsapp.content.read` juntos.

### Parâmetros de caminho

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `whatsapp_id` | `string` | Sim | — |

### Parâmetros de query

| Nome | Tipo | Obrigatório | Descrição |
| --- | --- | --- | --- |
| `limit` | `integer` | Não | Records per page, 1 to 100. A page may return fewer than this while has_more is still true; drive the loop by has_more, not by the size of data. |
| `cursor` | `string` | Não | Signed cursor returned by the previous page. |

### Exemplo de requisição

```bash
curl "https://api.salesbud.com.br/v1/whatsapp/mtg_01JEXAMPLE/messages?limit=...&cursor=..." \
  -H "Authorization: Bearer $SALESBUD_ACCESS_TOKEN"
```

### Exemplo de resposta

```json title="200 OK"
{
  "data": [
    {
      "id": "wamsg_01JEXAMPLE1",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:30:00.000Z",
      "direction": "inbound",
      "author": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "type": "text",
      "text": "Pode me mandar a proposta?",
      "media": null,
      "transcription": null,
      "edited_at": null,
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wamsg_01JEXAMPLE2",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:32:00.000Z",
      "direction": "outbound",
      "author": {
        "phone": "+5511988880001",
        "name": "Vendedora Exemplo"
      },
      "type": "audio",
      "text": null,
      "media": {
        "file_name": null,
        "mime_type": "audio/ogg; codecs=opus",
        "size_bytes": 48213
      },
      "transcription": "Claro, envio ainda hoje por e-mail.",
      "edited_at": null,
      "deleted_at": null,
      "deleted_reason": null
    },
    {
      "id": "wamsg_01JEXAMPLE3",
      "object": "whatsapp_message",
      "conversation_id": "wa_01JEXAMPLE",
      "sent_at": "2026-09-12T15:35:00.000Z",
      "direction": "inbound",
      "author": {
        "phone": "+5511999990001",
        "name": "Cliente Exemplo"
      },
      "type": "text",
      "text": null,
      "media": null,
      "transcription": null,
      "edited_at": null,
      "deleted_at": "2026-09-12T15:36:00.000Z",
      "deleted_reason": "revoked_for_everyone"
    }
  ],
  "pagination": {
    "limit": 50,
    "has_more": false,
    "next_cursor": null
  },
  "request_id": "req_ExampleRequestId01"
}
```

### Respostas

| Status | Descrição |
| --- | --- |
| `200` | Company-scoped WhatsApp message page. |
| `400` | Invalid request. |
| `401` | Missing or invalid credentials. |
| `403` | The token does not contain the required scope. |
| `404` | Resource not found in the authenticated company. |
| `429` | Client rate limit exceeded. |
| `503` | A required audit or rate-limit dependency is unavailable. |

:::note
Esta página é gerada a partir da especificação OpenAPI. Não edite à mão.
:::

---

# Changelog

Source: https://docs.salesbud.com.br/pt-br/resources/changelog

Mudanças que quebram uma versão já publicada são anunciadas antes de entrar.
Mudanças aditivas — campo novo, endpoint novo, código de erro novo — podem chegar
a qualquer momento, então faça parse defensivo: ignore campos que você não
reconhece.

## Não publicado — primeira entrega

A versão `v1` ainda não foi liberada para parceiros. O que segue é a forma da
primeira entrega.

**Ferramentas**

- Um servidor MCP oficial acompanha a v1: `@salesbud/mcp` no npm, uma tool por
  operação pública de leitura, para Claude Desktop, Claude Code e Cursor. Ele é
  cliente desta API e não acrescenta capacidade nenhuma. Veja
  [Servidor MCP](/pt-br/guides/mcp/).

**Recursos**

- Ligações são coleção própria em `/v1/calls`, com identificadores `call_`. Um
  registro capturado por integração de VoIP tem `object: "call"`; o resto é
  `meeting`. Ids não resolvem entre coleções.
- As cinco rotas são espelhadas nas duas coleções: listar, obter, transcrição,
  respostas de template e avaliação geral.
- A ligação não carrega `bot_history`: captura de VoIP não tem bot de gravação, então o campo é
  omitido em vez de voltar vazio, e um schema que valida ligação o recusa.
  `enablement.meeting_type` fica e costuma vir `null` — esse nulo é legítimo, diferente do bot.
- Conversas de WhatsApp dos vendedores da empresa são uma coleção em `/v1/whatsapp`,
  com identificadores `wa_` e mensagens `wamsg_`, atrás de `whatsapp.read` (metadado) e
  `whatsapp.content.read` (texto sanitizado, metadado de mídia, transcrição de áudio).
  `last_message_at` só anda com mensagem nova, então releia as mensagens para ver edições
  e apagamentos; mensagem apagada fica como lápide. Filtros de data comparam em segundos
  inteiros. Veja [WhatsApp](/pt-br/reference/whatsapp/listwhatsappconversations/).

**Erros**

- O envelope de erro é `{ type, title, detail, code, request_id }`. Todo campo é
  snake_case; o status HTTP fica na status line e não se repete no corpo.
  Ramifique pelo `code`.
- `POST /oauth/token` mantém o formato de erro do OAuth 2.0 exigido pela
  RFC 6749 §5.2.

**Paginação e filtros**

- Os cursores são assinados e ordenados por id crescente; eles amarram os filtros
  da requisição que os gerou.
- `updated_after`, `updated_before` e `snapshot_at` foram removidos. Sem carimbo
  dedicado de visibilidade não há sync incremental honesto; reconsulte por
  período. O `updated_at` continua nas respostas, como informativo.
- Os filtros de data são `meeting_after` e `meeting_before`. Versões anteriores também
  aceitavam `meeting_at_from` e `meeting_at_to`; eles nunca chegaram a parceiro e foram
  removidos, em vez de nascerem descontinuados.

**Acesso**

- A API é gateada pela feature `API_ACCESS` da empresa, verificada na emissão de
  token e na criação de credencial.
- Emails e domínios de conta são normalizados para minúsculo em todos os campos,
  então um endereço pode ser usado como chave de junção independentemente de onde
  apareça.

---

# Para LLMs e agentes

Source: https://docs.salesbud.com.br/pt-br/resources/for-agents

Esta documentação é publicada também em formato legível por máquina, para que um
agente a consuma sem raspar markup.

## Pontos de entrada

| Arquivo | Conteúdo |
| --- | --- |
| [`/llms.txt`](/llms.txt) | Índice: uma linha por página, com URL e resumo. Comece por aqui. |
| [`/llms-full.txt`](/llms-full.txt) | Todas as páginas concatenadas em Markdown puro, na ordem de leitura. |
| [`/openapi.yaml`](/openapi.yaml) | O contrato normativo. Tudo na referência da API é gerado dele. |

O formato segue a [convenção llms.txt](https://llmstxt.org/).

## Qualquer página em Markdown

Acrescente `.md` a qualquer URL da documentação para obter a fonte em vez da
página renderizada:

```
https://docs.salesbud.com.br/pt-br/guides/pagination/    → HTML
https://docs.salesbud.com.br/pt-br/guides/pagination.md  → Markdown
```

As duas versões de idioma estão disponíveis; as páginas em português ficam sob
`/pt-br/`.

## Se você só precisa ler dado a partir de um agente

Esta API serve um servidor MCP em `https://api.salesbud.com.br/mcp`, que o claude.ai
alcança como conector personalizado. Ele expõe todas as operações públicas de
leitura como tools e cuida de renovação de token, retries e janelamento de
transcrição. É o caminho mais curto entre uma credencial e um agente lendo
reuniões — veja [Servidor MCP](/pt-br/guides/mcp/).

Escreva um cliente HTTP quando estiver construindo um serviço, e não pilotando
um modelo.

## Se você vai gerar código de integração

Três coisas erram com facilidade e vale ler antes de escrever qualquer linha:

1. **Itere por `pagination.has_more`, nunca por `data.length`.** Uma página pode
   vir curta ou vazia com dado restante. Veja [Paginação](/pt-br/guides/pagination/).
2. **Ramifique por `error.code`, nunca por `error.detail`.** O detalhe é texto
   humano e pode ser reescrito. Veja [Erros](/pt-br/guides/errors/).
3. **Não deduza o tipo de recurso pelo `type`.** O `object` diz se é `meeting` ou
   `call`; o `type` diz só `video` ou `audio`, e os dois são independentes. Veja
   [Reuniões e ligações](/pt-br/guides/meetings-and-calls/).

Não existe refresh token — renovar é chamar `/oauth/token` de novo. Isso é
propriedade do fluxo client credentials, não omissão.