> ## Documentation Index > Fetch the complete documentation index at: https://docs.julian.11x.ai/llms.txt > Use this file to discover all available pages before exploring further. # Retrieve recordings for conversations > Retrieve recording URLs for every call in the requested conversation sequences, optionally with each call's transcript. Two limits apply. At most **100 sequence IDs** per request, and those sequences may contain at most **500 calls** in total — a few long sequences can exceed the call limit while well under the sequence limit, in which case split the IDs across requests. Nothing is truncated silently. Sequences you cannot see are reported in `notFound` rather than failing the request. ## OpenAPI ````yaml /openapi.json post /recordings openapi: 3.1.0 info: title: Julian API version: 1.0.0 description: >- API for Julian, 11x's AI calling platform. Programmatically enroll contacts into agent calling sequences, manage sequences, and track credit usage. All requests are authenticated with an organization API key in the `x-api-key` header and rate limited to 10 requests per second per key (HTTP 429 when exceeded). servers: - url: https://julian.11x.ai/api/v1 security: - apiKeyAuth: [] tags: - name: Scheduling description: Enroll contacts into an agent's calling sequence. - name: Sequences description: Manage active sequences. - name: Agents description: Read agent configuration. - name: Conversation Meetings description: >- Manage the meeting booked during a conversation — list its attendees, add attendees before it starts, and cancel it. Every endpoint is addressed by an `entityId` that identifies the conversation. Today only voice calls are supported, with the form of the call prefix followed by the call ID (e.g. `call.abc123`). The meeting is resolved from the calendar booking created during the conversation, and the calendar provider remains the source of truth. - name: Conversation Metadata description: >- Retrieve completed conversation data asynchronously — designed for batch ingestion, backfills, and reconciliation jobs that pull conversation metadata instead of receiving it through a webhook. **Recommended batch ingestion flow:** 1. Call `GET /conversations/search/{agentId}` with the time window you want to ingest (e.g. yesterday or the last seven days). 2. Iterate through the returned `sequenceIds`. 3. For each sequence ID, call `GET /conversations/metadata/{sequenceId}`. 4. Store the returned `metadata` array in your warehouse. 5. Use `limit`, `offset`, and `hasMore` to paginate through larger backfills. - name: Credits description: Track credit utilization. - name: Webhooks description: Events Julian sends to your endpoint. - name: Recordings description: >- Fetch call audio, and optionally transcripts. A recording belongs to a single call, so `GET /recordings/{callId}` is addressed by call ID. The bulk endpoint selects by **sequence** ID instead, because that is the unit `GET /conversations/search/{agentId}` returns — same resource, two selectors. **Bulk export flow:** 1. Call `GET /conversations/search/{agentId}` for the window you want. 2. Pass a page of the returned `sequenceIds` to `POST /recordings` (up to 100 per request). 3. Download each `url` before it expires; re-fetch individual calls later with `GET /recordings/{callId}`. Recording URLs are signed and short-lived — see `expiresAt` on each entry. paths: /recordings: post: tags: - Recordings summary: Retrieve recordings for conversations description: >- Retrieve recording URLs for every call in the requested conversation sequences, optionally with each call's transcript. Two limits apply. At most **100 sequence IDs** per request, and those sequences may contain at most **500 calls** in total — a few long sequences can exceed the call limit while well under the sequence limit, in which case split the IDs across requests. Nothing is truncated silently. Sequences you cannot see are reported in `notFound` rather than failing the request. operationId: listRecordings requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/RecordingsRequest' responses: '200': description: Recordings grouped by conversation sequence. content: application/json: schema: $ref: '#/components/schemas/RecordingsResponse' '400': description: >- `INVALID_REQUEST` — no sequence IDs, or more than 100. `TOO_MANY_CALLS` — the requested sequences contain more than 500 calls in total. `INVALID_REQUEST_BODY` — the body is not valid JSON. content: application/json: schema: $ref: '#/components/schemas/Error' '401': $ref: '#/components/responses/Unauthorized' '500': $ref: '#/components/responses/InternalServerError' components: schemas: RecordingsRequest: type: object properties: sequenceIds: type: array items: type: string minItems: 1 maxItems: 100 description: >- Conversation sequence IDs, as returned by `GET /conversations/search/{agentId}`. Duplicates are collapsed. example: - call-sequence.xyz123 - call-sequence.xyz456 includeTranscripts: type: boolean default: false description: >- When `true`, include each call's transcript text alongside its recording. required: - sequenceIds RecordingsResponse: type: object properties: recordings: type: array items: $ref: '#/components/schemas/SequenceRecordings' notFound: type: array items: type: string description: >- Requested sequence IDs that do not exist or do not belong to your organization. Omitted when every requested ID resolved. required: - recordings Error: type: object required: - error - message properties: error: type: string description: Machine-readable error code. example: INVALID_PHONE_NUMBER message: type: string description: Human-readable explanation. SequenceRecordings: type: object properties: sequenceId: type: string example: call-sequence.xyz123 calls: type: array description: One entry per call in the sequence, ordered by call step. items: $ref: '#/components/schemas/CallRecordingEntry' required: - sequenceId - calls CallRecordingEntry: type: object description: Recording — and optionally transcript — for a single call. properties: callId: type: string example: call.abc123 recordingStatus: type: string enum: - downloaded - no-recording - call-not-found description: >- Whether audio was located for this call. Describes the **audio only** — a call can have a transcript and no recording, so do not filter on this value when you also want transcripts. - `downloaded` — audio was located and `url` is present. - `no-recording` — the call has no retrievable audio (never recorded, cleaned up, or hosted by a provider this API cannot sign for). - `call-not-found` — no call exists for the id. url: type: string format: uri description: >- Short-lived link to the call audio. Absent when `recordingStatus` is not `downloaded`. Treat this as a secret: anyone holding the URL can fetch the audio until it expires, so do not forward it into tickets or third-party tools. expiresAt: type: string format: date-time description: >- When `url` stops working — **3 hours** after the response is generated. Present only for URLs this API signs; provider-hosted audio carries its own lifetime, which we do not control and therefore do not report. Re-fetch with `GET /recordings/{callId}` at any time. durationSeconds: type: integer example: 184 scheduledAt: type: string format: date-time stepNumber: type: integer description: Position of this call within its sequence. example: 1 transcript: type: string description: >- Formatted transcript text. Present only when transcripts were requested and the call has one — independent of `recordingStatus`. example: '[Agent] (0:00): Hi, is this Victor?' required: - callId - recordingStatus responses: Unauthorized: description: >- `UNAUTHORIZED` or `INVALID_API_KEY` — API key missing, invalid, or not authorized for this agent's organization. content: application/json: schema: $ref: '#/components/schemas/Error' examples: missingApiKey: summary: Missing API key value: error: UNAUTHORIZED message: >- API key is required. Please provide your API key in the x-api-key header. invalidApiKey: summary: Invalid API key value: error: INVALID_API_KEY message: The provided API key is invalid. InternalServerError: description: >- `INTERNAL_SERVER_ERROR` — unexpected server error. Retry with backoff and contact the 11x team if it persists. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: INTERNAL_SERVER_ERROR message: Internal Server Error. Please contact 11x team. securitySchemes: apiKeyAuth: type: apiKey in: header name: x-api-key description: Organization API key. Get it from the 11x team or from the platform. ````