generated: '2026-08-05'
method: searched
source: https://support.read.ai/hc/en-us/articles/49381161088659-API-Reference
description: >-
  Cross-cutting request/response semantics for the Read AI REST API v1, read from the
  provider's own API Reference, auth guide and webhook guide. Read AI has no OpenAPI,
  so every convention below is captured from prose + published examples.
authentication:
  style: OAuth 2.1 bearer token in the Authorization header
  header: 'Authorization: Bearer <access_token>'
  detail: authentication/read-ai-authentication.yml
  static_keys: false
base_url: https://api.read.ai/
versioning:
  scheme: uri-path
  current: v1
  maturity: open beta
  detail: lifecycle/read-ai-lifecycle.yml
idempotency:
  supported: false
  note: >-
    Read AI documents no idempotency key, no request-replay contract and no unsafe
    methods on the public API — every documented v1 endpoint is a GET, so idempotency
    is a property of the verb rather than a published contract. The one replay-safety
    mechanism Read AI does ship is on the OUTBOUND side: every webhook delivery
    carries a unique request_id the receiver can use to discard duplicates. No
    Idempotency pointer is emitted for this provider.
  outbound_replay_protection:
    field: request_id
    where: webhook payload body
    purpose: detect and ignore duplicate deliveries, preventing replay attacks
pagination:
  style: cursor
  parameters:
    - name: limit
      description: Maximum objects returned. Default 10, max 10.
    - name: cursor
      description: >-
        ID of the last object in the previous page's data array. Set it to continue
        after that object.
  response_fields:
    - {name: object, type: string, description: 'Typically "list".'}
    - {name: url, type: string, description: URL of the request.}
    - {name: has_more, type: boolean, description: Whether additional pages exist.}
    - {name: data, type: array, description: The page of meeting objects.}
  ordering: reverse chronological (newest first)
  note: >-
    The MCP tool list_meetings inherits the same 10-item cap and cursor contract.
field_expansion:
  parameter: 'expand[]'
  style: repeated query parameter, e.g. ?expand[]=summary&expand[]=metrics
  fields:
    - summary
    - chapter_summaries
    - action_items
    - key_questions
    - topics
    - transcript
    - metrics
    - recording_download
  live_meeting_fields: [transcript, chapter_summaries]
  cost_note: >-
    Read AI warns that expansion increases response time, and that expanding multiple
    fields on a list request can be noticeably slower.
filtering:
  style: dotted comparison operators on a field name
  parameters:
    - 'start_time_ms.gt'
    - 'start_time_ms.gte'
    - 'start_time_ms.lt'
    - 'start_time_ms.lte'
  time_format: Unix epoch milliseconds (integer)
identifiers:
  meeting_id: ULID
  example_shape: 26-character Crockford base32 ULID
  transcript_timestamps: Unix epoch milliseconds
  webhook_times: ISO 8601 UTC (RFC 3339)
error_envelope:
  format: plain HTTP status codes with a JSON body
  observed_body: '{"detail": "<message>"}'
  observed_from: >-
    Anonymous probes of https://api.read.ai/v1/meetings (401) and unknown paths (404)
    on 2026-08-05; the shape is FastAPI's default detail envelope.
  problem_json: false
  detail: errors/read-ai-problem-types.yml
rate_limiting:
  limit: 100 requests per minute per user
  exceeded_status: 429
  headers_documented: false
  detail: rate-limits/read-ai-rate-limits.yml
request_tracing:
  inbound_request_id_header: null
  note: >-
    No inbound request-id/correlation header is documented. Outbound webhooks carry a
    request_id in the body.
webhooks:
  signature_header: X-Read-Signature
  algorithm: HMAC-SHA256 over the raw request body, key base64-decoded
  detail: asyncapi/read-ai-webhooks.yml
content_type:
  request: application/json
  response: application/json
  accept_header_used_in_docs: application/json
cross_links:
  authentication: authentication/read-ai-authentication.yml
  scopes: scopes/read-ai-scopes.yml
  errors: errors/read-ai-problem-types.yml
  lifecycle: lifecycle/read-ai-lifecycle.yml
  rate_limits: rate-limits/read-ai-rate-limits.yml
  webhooks: asyncapi/read-ai-webhooks.yml
  data_model: data-model/read-ai-data-model.yml