> ## Documentation Index > Fetch the complete documentation index at: https://docs.octavehq.com/llms.txt > Use this file to discover all available pages before exploring further. # Search Call Transcripts > Semantic search over recorded call transcripts, returning verbatim, speaker-attributed moments with recording links and time anchors. Filter by company, library entity (competitor/objection/persona/...), speaker side (customer voice vs rep talk track), specific speaker or persona cohort (who SAID it), deal outcome (WON/LOST/OPEN, resolved live from CRM), a single call or opportunity, call sentiment, and date range. Calls are grouped with their linked CRM opportunities and current stage, so results can be read as won-deal or lost-deal precedents. ## OpenAPI ````yaml post /api/v2/transcript/search openapi: 3.0.0 info: version: 5.4.0 title: Octave API description: API for Octave workspace management and AI-powered content generation servers: - url: https://app.octavehq.com security: - ApiKeyAuth: [] paths: /api/v2/transcript/search: post: tags: - Transcript summary: Search Call Transcripts description: >- Semantic search over recorded call transcripts, returning verbatim, speaker-attributed moments with recording links and time anchors. Filter by company, library entity (competitor/objection/persona/...), speaker side (customer voice vs rep talk track), specific speaker or persona cohort (who SAID it), deal outcome (WON/LOST/OPEN, resolved live from CRM), a single call or opportunity, call sentiment, and date range. Calls are grouped with their linked CRM opportunities and current stage, so results can be read as won-deal or lost-deal precedents. operationId: searchTranscripts requestBody: description: Transcript search query and filters content: application/json: schema: type: object properties: query: type: string minLength: 1 description: >- What to find in call transcripts, e.g. 'objections about pricing' or 'how prospects describe onboarding pain' example: objections about pricing companyDomain: type: string description: >- Restrict to calls with one company by its website domain (e.g. 'acme.com') example: acme.com entityOId: type: string description: >- Restrict to moments linked to a library entity (competitor, objection, persona, etc.) by its oId speakerSide: type: string enum: - external - internal - any default: any description: >- 'external' finds customer/prospect voice, 'internal' finds rep talk tracks speakerEmail: type: string description: Only moments where this person spoke (contact email) attributedPersonaOIds: type: array items: type: string description: >- Only moments spoken BY contacts classified into any of these personas, e.g. 'objections raised by CTOs or CISOs'. Different from entityOId, which finds moments ABOUT an entity attributedSegmentOIds: type: array items: type: string description: >- Only calls with companies classified into any of these segments — quotes from within a market segment, resolved live from company classifications eventOIds: type: array items: type: string description: >- Restrict to specific calls (Event oIds) — chain from /event/list results contentFilter: $ref: '#/components/schemas/TranscriptSearchContentFilter' opportunityId: type: string description: >- Restrict to calls linked to one CRM opportunity/deal. Accepts either an Octave oId (crmo_*) or the CRM's native id (Salesforce Id / HubSpot object id / Attio record UUID, e.g. pasted from a CRM URL) — pass whichever you have dealOutcome: type: string enum: - WON - LOST - OPEN description: >- Restrict to calls on deals with this outcome (resolved live from CRM), e.g. 'objection quotes on deals we lost' sentiment: type: string enum: - POSITIVE - NEUTRAL - NEGATIVE - UNKNOWN description: Only calls with this overall pipeline-extracted sentiment callPurposes: type: array items: type: string enum: - DISCOVERY - DEMO - TECHNICAL_EVALUATION - WORKING_SESSION - NEGOTIATION_PRICING - ONBOARDING - CHECK_IN_SUCCESS - QBR_RENEWAL - EXPANSION_UPSELL - ESCALATION_CHURN_RISK - SUPPORT - INTERNAL_SYNC - OTHER description: >- Only calls classified with any of these purposes (OR). E.g. ['ESCALATION_CHURN_RISK'] for save-attempt voice, ['DISCOVERY','DEMO'] for early-funnel calls expandQuery: type: boolean default: false description: >- Set true when the query is a short or abstract topic (a bare entity name, 'pricing pushback') — it gets rewritten into a hypothetical transcript passage before matching, improving recall. Leave false when the query already reads like something a speaker would say verbatim. Adds ~1s latency minRelevance: type: number minimum: 0 maximum: 1 description: >- Relative relevance floor (0-1): drop moments scoring below this fraction of the BEST match's score. Use ~0.5-0.7 when fewer, higher-confidence quotes beat volume; omit for exploratory searches preferRecent: type: boolean default: false description: >- Decay scores by call age (90-day half-life) so newer calls win ties. Use for 'current objections', 'latest messaging', trend questions; leave false for all-time or historical evidence. For a hard cutoff use sinceDate instead alpha: type: number minimum: 0 maximum: 1 description: >- Semantic-vs-keyword weight (default 0.7). Lower toward 0.3 when exact words must appear verbatim (product names, jargon, quoted phrases); raise toward 0.9 for conceptual/paraphrase matching where wording will differ sinceDate: type: string description: Only calls on or after this ISO date (e.g. 2026-04-01) example: '2026-04-01' untilDate: type: string description: >- Only calls on or before this ISO date (a date-only value is inclusive of the whole day it names) includeInternalMeetings: type: boolean default: false description: >- Also search calls with no external participant (internal meetings) limit: type: integer minimum: 1 maximum: 25 default: 8 description: 'Maximum number of moments to return (default: 8, max: 25)' example: 8 required: - query responses: '200': description: Matching calls with their moments content: application/json: schema: type: object properties: _metadata: $ref: '#/components/schemas/Metadata' query: type: string calls: type: array items: $ref: '#/components/schemas/TranscriptSearchCall' description: Matching calls with their moments totalMoments: type: number latestCallAt: type: string nullable: true description: Timestamp of the workspace's most recent processed call indexedThrough: type: string nullable: true description: >- Timestamp of the newest call present in the transcript index. When this is older than latestCallAt (or null), the newest call(s) are not searchable yet — results may miss them. required: - _metadata - query - calls - totalMoments - latestCallAt - indexedThrough '400': description: Bad request content: application/json: schema: type: object properties: _metadata: $ref: '#/components/schemas/Metadata' message: type: string required: - _metadata - message deprecated: false components: schemas: TranscriptSearchContentFilter: type: object properties: momentPhrases: type: array items: type: string minLength: 2 maxItems: 10 description: >- Literal, case-insensitive phrases — the returned QUOTE must contain at least one. Use when the exact words must be in the excerpt, e.g. ['rip and replace'] callPhrases: type: array items: type: string minLength: 2 maxItems: 10 description: >- Literal phrases the CALL must mention anywhere — the query then searches within those calls; the returned moment need not contain them. E.g. phrases ['Loops'] + query 'objections raised' finds objections on calls where Loops came up excludeCallPhrases: type: array items: type: string minLength: 2 maxItems: 10 description: Calls mentioning any of these anywhere are excluded entirely description: >- Exact-phrase content criteria. Within each list ANY phrase matches (OR — list synonyms); across fields all conditions apply (AND) Metadata: type: object properties: usage: type: number default: 0 example: 0 description: API usage requestId: type: string example: requestId description: Request ID message: type: string example: message description: Message timestamp: type: string example: '2021-01-01T00:00:00.000Z' description: Timestamp required: - requestId - timestamp TranscriptSearchCall: type: object properties: eventOId: type: string description: Event oId of the call callTitle: type: string nullable: true companyName: type: string nullable: true callDate: type: string nullable: true description: ISO date of the call callPurpose: type: string nullable: true recordingUrl: type: string nullable: true description: Recording link; combine with startSec to jump to the moment opportunities: type: array items: type: object properties: oId: type: string name: type: string stageCategory: type: string nullable: true description: Current stage category (queried live, e.g. WON/LOST) required: - oId - name description: CRM opportunities linked to this call moments: type: array items: $ref: '#/components/schemas/TranscriptSearchMoment' required: - eventOId - opportunities - moments TranscriptSearchMoment: type: object properties: quote: type: string description: Verbatim, speaker-attributed transcript excerpt startSec: type: number nullable: true description: Seconds into the call where the moment starts speakerNames: type: array items: type: string description: Speakers in this moment extractionTypes: type: array items: type: string description: Finding types the pipeline tagged this moment with score: type: number nullable: true description: Relevance score required: - quote - speakerNames - extractionTypes securitySchemes: ApiKeyAuth: type: apiKey in: header name: api_key ````