Get a single chat

Retrieves a specific chat by its ID, including its full transcript and any pending confirmations.

GET https://chat.betterstack.com/api/v2/chats/{id}

URL parameters

required string

The ID of the chat to retrieve.

Headers

Authorization

required string

Bearer $TOKEN

200

Returns the requested chat with its transcript and pending confirmations.

Response body

{
  "data": {
    "id": "12345",
    "type": "chat",
    "attributes": {
      "team_id": 123456,
      "team_name": "example-team",
      "name": "Checkout 5xx spike",
      "status": "idle",
      "message_count": 2,
      "last_message_at": "2026-09-07T10:32:18.000Z",
      "total_tokens": 15420,
      "model": "anthropic/claude-opus-5.5",
      "total_cost_usd": 0.0928,
      "created_at": "2026-09-07T10:30:00.000Z",
      "updated_at": "2026-09-07T10:32:18.000Z",
      "messages": [
        {
          "id": 9001,
          "message_type": "user",
          "text": "Why did checkout start returning 5xx at 10:15?",
          "artifacts": [],
          "created_at": "2026-09-07T10:30:00.000Z",
          "user": null
        },
        {
          "id": 9002,
          "message_type": "ai",
          "text": "The 5xx responses come from checkout-api pods restarting after the 10:12 deploy of v2.4.0. The new version fails its readiness probe because PAYMENTS_URL is unset. Set it in the deployment manifest and roll out again.",
          "artifacts": [
            {
              "type": "report",
              "title": "Checkout 5xx investigation",
              "content": "Summary: the deploy of v2.4.0 at 10:12 dropped the PAYMENTS_URL variable from the checkout-api deployment. Pods fail their readiness probe and restart in a loop. Setting the variable and rolling out again restores checkout."
            }
          ],
          "created_at": "2026-09-07T10:32:18.000Z",
          "user": null
        }
      ],
      "pending_confirmations": []
    }
  }
}

401

Unauthorized

Response body

{
  "errors": "Invalid authentication"
}

404

Not Found

Response body

{
  "errors": "Resource with the provided ID was not found"
}

Response attributes

string

The ID of the chat.

type

string

Always chat.

attributes.team_id

integer

The ID of the team the chat belongs to.

attributes.team_name

string

The name of the team the chat belongs to.

attributes.name

string

The name of the chat.

attributes.status

string

Whether the agent is working on the chat: processing, awaiting_confirmation, or idle. Chat statuses explains each.

attributes.message_count

integer

The number of messages in the chat.

attributes.last_message_at

string or null

When the latest message was posted, as an ISO 8601 date-time in UTC, or null when the chat has no messages.

attributes.total_tokens

integer

The number of tokens the agent has used on the chat so far.

attributes.model

string or null

The model the chat runs on, such as anthropic/claude-opus-5.5, or null on chats created before model selection.

attributes.total_cost_usd

number

What the chat has cost so far, in US dollars.

attributes.created_at

string

When the chat was created, as an ISO 8601 date-time in UTC.

attributes.updated_at

string

When the chat was last updated, as an ISO 8601 date-time in UTC.

attributes.messages

array

The messages of the chat, oldest first.

attributes.messages[].id

integer

The ID of the message.

attributes.messages[].message_type

string

Who wrote the message: user or ai.

attributes.messages[].text

string or null

The text of the message as markdown, which on an AI message is the answer and is null until the agent produces one.

attributes.messages[].artifacts

array

The structured content of the message: the agent's reports on an AI message, or the files and references a user attached.

attributes.messages[].artifacts[].type

string

The kind of artifact: report on AI messages, or file, incident, error, text_blob, investigation, prompt, or live_tail_filter on user messages.

attributes.messages[].artifacts[].title

string

The title of a report, investigation, or prompt, or the source names of a live_tail_filter.

attributes.messages[].artifacts[].content

string

The text of a report as markdown, or the pasted text of a text_blob.

attributes.messages[].artifacts[].name

string

The file name of a file, or the name of an incident or text_blob.

attributes.messages[].artifacts[].url

string

A link to the file, incident, or error.

attributes.messages[].artifacts[].size

integer

The size of a file in bytes.

attributes.messages[].artifacts[].content_type

string

The MIME type of a file, such as image/png.

attributes.messages[].artifacts[].id

integer

The ID of an incident.

attributes.messages[].artifacts[].error_type

string

The exception type of an error.

attributes.messages[].artifacts[].error_message

string

The exception message of an error.

attributes.messages[].artifacts[].source_id

integer

The ID of the application an error belongs to, or of the source an investigation looks into.

attributes.messages[].artifacts[].pattern

string

The pattern hash that identifies an error.

attributes.messages[].artifacts[].resource_type

string

What an investigation looks into: log, span, error, or exception.

attributes.messages[].artifacts[].dt

string

The timestamp of the log or span row an investigation looks into.

attributes.messages[].artifacts[].trace_id

string

The trace ID of the span an investigation looks into.

attributes.messages[].artifacts[].span_id

string

The span ID of the span an investigation looks into.

attributes.messages[].artifacts[].configuration_name

string

The name of the Uptime AI SRE integration a prompt comes from.

attributes.messages[].artifacts[].source_ids

string

The comma-separated IDs of the sources a live_tail_filter covers.

attributes.messages[].artifacts[].query

string

The filter query of a live_tail_filter.

attributes.messages[].artifacts[].from

string

The start of a live_tail_filter's time range, absolute or relative like now-3h.

attributes.messages[].artifacts[].to

string

The end of a live_tail_filter's time range, absolute or relative like now.

attributes.messages[].artifacts[].range

string

The label of a live_tail_filter's time range, such as Last 3 hours.

attributes.messages[].created_at

string

When the message was posted, as an ISO 8601 date-time in UTC.

attributes.messages[].user

object or null

The user who wrote the message, or null when no user is behind it, as on AI messages and messages sent with a team API token.

attributes.messages[].user.better_stack_id

integer

The Better Stack ID of the user.

attributes.messages[].user.name

string

The full name of the user, or their e-mail when no name is set.

attributes.pending_confirmations

array

The tool runs waiting for your approval, oldest first, empty unless the status is awaiting_confirmation.

attributes.pending_confirmations[].id

integer

The ID of the confirmation, used to approve or deny it.

attributes.pending_confirmations[].tool_name

string

The name of the tool the agent wants to run, such as Bash.

attributes.pending_confirmations[].tool_input

object

The input the agent wants to run the tool with, such as the command.

attributes.pending_confirmations[].status

string

Always pending, since only confirmations still waiting for a decision are listed.

attributes.pending_confirmations[].resolution_note

string or null

The note left with the decision, always null here since the confirmation is still pending.

attributes.pending_confirmations[].expires_at

string

When the confirmation expires if nobody approves or denies it, as an ISO 8601 date-time in UTC.

attributes.pending_confirmations[].confirmed_at

string or null

When the confirmation was approved or denied, always null here since the confirmation is still pending.

attributes.pending_confirmations[].created_at

string

When the agent asked for the confirmation, as an ISO 8601 date-time in UTC.

Explore documentation

Example cURL

Explore documentation

Get a single chat

Example cURL

Please accept cookies