> ## Documentation Index > Fetch the complete documentation index at: https://docs.support.onclarity.com/llms.txt > Use this file to discover all available pages before exploring further. # Get conversation artifacts > Returns everything recorded about a finished conversation: for each agent session, the transcript, the event timeline, detected problems (`findings`), per-turn latency, and optionally a download link for the audio recording; plus a `merged` view that joins all sessions in time order. This is the endpoint to use for transcripts and recordings. Artifacts are uploaded shortly after a call ends. Until they arrive — and while personal data is still being redacted from them, on organizations that have redaction switched on — the endpoint answers 404 with `code: ARTIFACTS_NOT_READY`; retry after a short wait. When only some sessions could be loaded, the response is 200 and the missing sessions carry an `error` object instead of their artifacts. ## OpenAPI ````yaml /api-reference/voice-openapi.json get /v1/projects/{projectId}/conversations/{id}/artifacts openapi: 3.0.3 info: title: Clarity Voice API version: '1.0' description: Voice agents, outbound calls, conversations, campaigns, and evals. servers: - url: https://voice-api.onclarity.com security: - organizationApiKey: [] tags: - name: Projects description: >- The projects your organization owns. Every other resource lives in a project. - name: Agents description: Create, configure, version and publish voice agents. - name: Outbound calls description: Have a published agent phone a number now. - name: Conversations description: >- One record per call session, with its calls, recordings, transcripts and artifacts. - name: Calls description: The calls of a project, with their status, direction, timing and outcome. - name: Campaigns description: Outbound calling campaigns that phone a list of numbers on a schedule. - name: Eval suites description: >- Reusable sets of eval scenarios and test cases to measure an agent against. - name: Eval judges description: The LLM judges that score eval calls against your test cases. - name: Eval scenarios description: The simulated-caller scenarios of your eval suites. - name: Eval test cases description: The checks a judge applies to each eval call. - name: Eval submissions description: Run an eval suite against an agent and read the verdicts. paths: /v1/projects/{projectId}/conversations/{id}/artifacts: parameters: - $ref: '#/components/parameters/ProjectIdPath' - $ref: '#/components/parameters/ConversationId' get: tags: - Conversations summary: Get conversation artifacts description: >- Returns everything recorded about a finished conversation: for each agent session, the transcript, the event timeline, detected problems (`findings`), per-turn latency, and optionally a download link for the audio recording; plus a `merged` view that joins all sessions in time order. This is the endpoint to use for transcripts and recordings. Artifacts are uploaded shortly after a call ends. Until they arrive — and while personal data is still being redacted from them, on organizations that have redaction switched on — the endpoint answers 404 with `code: ARTIFACTS_NOT_READY`; retry after a short wait. When only some sessions could be loaded, the response is 200 and the missing sessions carry an `error` object instead of their artifacts. operationId: getConversationArtifacts parameters: - name: include_recording in: query required: false schema: type: string enum: - 'true' - 'false' default: 'false' description: >- `true` adds a short-lived signed download link for each session's audio recording (`sessions[].recording_signed_url`, expiring at `recording_signed_url_expires_at`). - name: include_session_config in: query required: false schema: type: string enum: - 'true' - 'false' default: 'false' description: >- `true` adds each session's effective agent configuration as it ran (`sessions[].session_config`), when it was stored. responses: '200': description: The conversation's artifacts. content: application/json: schema: $ref: '#/components/schemas/ConversationArtifacts' example: project_id: 3c59dc04-8a4b-4f7e-9b1a-6c2a0e5d1f22 conversation_id: 8f14e45f-ceea-467f-a0e6-3f1c2b9d7a10 bucket: example-bucket sessions_count: 1 sessions_degraded: 0 sessions: - session_id: AJ_x8Kd2mQp4Vn call_id: 1b6453892473a467d07372d45eb05abc agent_id: 6f4922f4-5568-461a-8f1d-9c3b2a1e7d55 joined_at: '2026-10-20T09:15:02.000Z' left_at: '2026-10-20T09:18:06.000Z' gcs_prefix: >- gs://example-bucket/calls/3c59dc04-8a4b-4f7e-9b1a-6c2a0e5d1f22/AJ_x8Kd2mQp4Vn/ transcript: |- AGENT: Hello, how can I help? CALLER: I'd like to move my appointment to Thursday. findings: [] fatal_count: 0 warning_count: 0 recording_signed_url: >- https://storage.googleapis.com/example-bucket/calls/3c59dc04-8a4b-4f7e-9b1a-6c2a0e5d1f22/AJ_x8Kd2mQp4Vn/recording.wav?X-Goog-Signature=abc123 recording_signed_url_expires_at: '2026-10-20T10:18:06.000Z' merged: transcript: |- AGENT: Hello, how can I help? CALLER: I'd like to move my appointment to Thursday. findings: [] fatal_count: 0 warning_count: 0 media_insights: [] messages: [] '401': $ref: '#/components/responses/Unauthorized' '403': $ref: '#/components/responses/Forbidden' '404': description: >- The conversation does not exist, or its artifacts are not available yet (`code: ARTIFACTS_NOT_READY` — retry later) or were never stored (`code: NO_ARTIFACTS`, e.g. a call nobody joined). content: application/json: schema: $ref: '#/components/schemas/Error' example: error: >- Artifacts not ready for conversation 8f14e45f-ceea-467f-a0e6-3f1c2b9d7a10 — all 1 session(s) reported ARTIFACTS_NOT_READY (transient upload lag or an aborted call that stored nothing). code: ARTIFACTS_NOT_READY '500': description: Recording storage is not configured on this deployment. content: application/json: schema: $ref: '#/components/schemas/Error' '502': description: >- Storage failed for every session of the conversation (`code: OBJECT_STORAGE_ERROR`). Retry later. content: application/json: schema: $ref: '#/components/schemas/Error' components: parameters: ProjectIdPath: name: projectId in: path required: true schema: type: string description: Project id ConversationId: name: id in: path required: true schema: type: string description: Conversation id schemas: ConversationArtifacts: type: object required: - project_id - conversation_id - bucket - sessions_count - sessions_degraded - sessions - merged properties: project_id: type: string conversation_id: type: string bucket: type: string example: example-bucket sessions_count: type: integer minimum: 0 sessions_degraded: type: integer minimum: 0 description: Count of sessions that failed to load (have an `error` field). sessions: type: array items: $ref: '#/components/schemas/SessionArtifacts' merged: $ref: '#/components/schemas/MergedArtifacts' media_insights: type: array description: >- Caller-side network stats per call (WebRTC RTT/jitter/loss measured at the caller's browser), collected across the conversation's calls. Empty for SIP-only conversations and calls where the widget never reported. items: type: object additionalProperties: true properties: call_id: type: string conversation_id: type: string nullable: true messages: type: array description: >- In-call WhatsApp / email sends of this conversation with their events (statuses, button taps, replies, no-response notes), including events that arrived after hang-up. The recipient is masked; reply texts are withheld on calls with PII redaction on. items: type: object properties: id: type: string channel: type: string enum: - whatsapp - email kind: type: string enum: - template - text - document - buttons - email message_ref: type: string recipient: type: string nullable: true recipient_source: type: string nullable: true status: type: string failure_reason: type: string nullable: true origin: type: string created_at: type: string format: date-time events: type: array items: type: object properties: event_type: type: string status: type: string nullable: true failure_reason: type: string nullable: true button_id: type: string nullable: true button_label: type: string nullable: true text: type: string nullable: true provider_ts: type: string nullable: true received_at: type: string delivered_to_agent: type: boolean Error: type: object required: - error additionalProperties: true properties: error: type: string description: Error message code: type: string description: >- A stable machine-readable reason, present on the refusals a client may branch on. Among them: `PUBLISH_CAPABILITY_REQUIRED` (403, the credential may not change what is live), `AGENT_NOT_PUBLISHED` (409, the agent has no published version yet), `AGENT_TYPE_NOT_DISPATCHABLE` (422, a Voice Assist agent named where an agent must place or take a call; the body carries `agent_type` and `agent_id`), `NOT_APPLICABLE_TO_AGENT_TYPE` (422, a voice-only setting written on a Voice Assist agent), `AGENT_TYPE_BOUNDARY` (409), `AGENT_TYPE_NOT_SUPPORTED` (400), `VOICE_ASSISTANT_REQUIRED` (403, the organization is not licensed for Voice Assist), `ORGANIZATION_UNAVAILABLE` (503) and `REGION_PROFILE_CONFLICT` (503). SessionArtifacts: type: object description: | One agent session's artifacts. On success, the artifact fields (metadata, timeline_jsonl, transcript, findings, summary, counts, recording_*) are populated. On failure, `error` is populated and the artifact fields are absent. required: - session_id - call_id - agent_id - joined_at - left_at - gcs_prefix properties: session_id: type: string description: Agent session id — the GCS prefix key. Authoritative. call_id: type: string nullable: true description: >- DB call row whose `participant_identity == session_id`, if any. Display-only. agent_id: type: string nullable: true description: | The agent that ran the session; null once that agent is deleted. A deleted agent's sessions are still served, with their transcript and recording. joined_at: type: string format: date-time nullable: true left_at: type: string format: date-time nullable: true gcs_prefix: type: string example: gs://example-bucket/calls/// metadata: additionalProperties: true type: object timeline_jsonl: type: string description: >- Raw NDJSON. One event per line, 1-based line numbers cross-referenced by `findings[].evidence.line`. transcript: type: string findings: type: array items: $ref: '#/components/schemas/TimelineFinding' summary: $ref: '#/components/schemas/TimelineSummary' fatal_count: type: integer minimum: 0 warning_count: type: integer minimum: 0 recording_signed_url: type: string format: uri nullable: true recording_signed_url_expires_at: type: string format: date-time nullable: true dictionary_usage: allOf: - $ref: '#/components/schemas/DictionaryUsage' nullable: true description: > Dictionary entries in effect for this session, whether each kind could take effect, and per-entry hits with the transcript lines they touched. `null` when the session recorded none (calls before 088, runtimes that do not record it, degraded legs) — never zero counts. caller_auth: $ref: '#/components/schemas/CallerAuthSummary' error: type: object description: >- Present instead of the artifact fields when this session's GCS fetch failed. required: - status - message properties: status: type: integer message: type: string MergedArtifacts: type: object description: >- Cross-session rollup. Findings are tagged with `session_id` so callers can locate the source session. required: - transcript - findings - summary - fatal_count - warning_count properties: transcript: type: string description: Chronologically-merged transcript across all sessions. findings: type: array items: allOf: - $ref: '#/components/schemas/TimelineFinding' - type: object required: - session_id properties: session_id: type: string summary: $ref: '#/components/schemas/TimelineSummary' fatal_count: type: integer minimum: 0 warning_count: type: integer minimum: 0 TimelineFinding: type: object description: >- Deterministic classifier finding over `timeline_jsonl`. `evidence.line` is a 1-based line number callers can Read directly to validate the diagnosis. required: - category - severity - summary properties: category: type: string enum: - error_event - log_error - tool_failure - silence_recovery - repeated_nudges - unexpected_end - tool_unfinished - malformed_event severity: type: string enum: - fatal - warning - info summary: type: string evidence: type: object properties: line: minimum: 1 type: integer lines: type: array items: minimum: 1 type: integer ts: type: number snippet: type: string extra: additionalProperties: true type: object suggested_file: type: string TimelineSummary: type: object description: Aggregate shape of a session's timeline. required: - total_events - session_started - session_ended - end_call_requested - nudge_count properties: total_events: type: integer minimum: 0 session_started: type: boolean session_ended: type: boolean end_call_requested: type: boolean nudge_count: type: integer minimum: 0 first_ts: type: number nullable: true last_ts: type: number nullable: true DictionaryUsage: type: object description: >- Which dictionary entries were in effect for the session and how they were used. required: - version - kinds - totals - tier_switched properties: version: type: integer example: 1 kinds: type: array description: Always three, in order replacements, keyterms, pronunciations. items: $ref: '#/components/schemas/DictionaryKindUsage' totals: type: object required: - entries - used - hits properties: entries: type: integer minimum: 0 used: type: integer minimum: 0 hits: type: integer minimum: 0 tier_switched: type: boolean description: >- The STT or TTS tier was not the primary at call end; the status describes the primary. CallerAuthSummary: type: object nullable: true description: >- How caller authentication went on one session. Null when the session recorded none. Never carries a private session value. required: - enabled - outcome - attempts_used - max_attempts - verified_at - identities - refused_tools - exhausted_action properties: enabled: type: boolean enum: - true outcome: type: string enum: - verified - not_verified - exhausted attempts_used: type: integer minimum: 0 max_attempts: type: integer nullable: true verified_at: type: number nullable: true description: Unix epoch seconds. identities: type: array items: additionalProperties: type: string type: object refused_tools: type: array items: type: object required: - tool - ts properties: tool: type: string ts: type: number nullable: true description: Unix epoch seconds. exhausted_action: type: string nullable: true enum: - end_call - transfer - continue - null DictionaryKindUsage: type: object required: - kind - state - reason - provider - model - strategy - mode - hit_label - entries properties: kind: type: string enum: - replacements - keyterms - pronunciations state: type: string description: >- active | not_supported | disabled | not_applicable (readers MUST tolerate other values). reason: type: string nullable: true description: >- stt_provider_ignores_keyterms | tts_provider_ignores_pronunciations | tts_fallback_mixes_providers | pronunciations_not_built | realtime_model | no_tts_provider (readers MUST tolerate other values). provider: type: string nullable: true model: type: string nullable: true strategy: type: string nullable: true description: 'Replacements only: manual | provider_api.' mode: type: string nullable: true description: 'Pronunciations only: text | ssml.' hit_label: type: string enum: - applied - heard - found_in_result entries: type: array description: Sorted by hits desc, then runtime order. items: $ref: '#/components/schemas/DictionaryEntryUsage' DictionaryEntryUsage: type: object required: - index - dictionary - hits - trackable - occurrences - overrides properties: index: type: integer minimum: 0 find: type: string replace: type: string term: type: string pronunciation: type: string dictionary: type: object nullable: true required: - id - name properties: id: type: string name: type: string hits: type: integer minimum: 0 trackable: type: boolean description: >- false for a provider-applied deletion, whose hits cannot be found in the result. occurrences: type: array items: type: object required: - line - line_ts - speaker - count - derived properties: line: type: integer description: >- 1-based line in this session's timeline_jsonl — the transcript row's `(Lnnn)` tag. line_ts: type: number nullable: true speaker: type: string enum: - caller - agent count: minimum: 1 type: integer derived: type: boolean overrides: type: array items: type: object required: - dictionary - text properties: dictionary: type: object required: - id - name properties: id: type: string name: type: string text: type: string responses: Unauthorized: description: No credential was sent, or the API key is invalid or revoked. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: Invalid or revoked token Forbidden: description: >- The credential carries no access to this project, or the operation needs administrative rights content: application/json: schema: $ref: '#/components/schemas/Error' securitySchemes: organizationApiKey: type: http scheme: bearer description: >- Your organization API key (it starts with `anek_org_`). The same key works on the Voice API and the Support API. See [Authentication](/authentication). ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.