openapi: 3.2.0 info: contact: email: help@avoma.com name: Avoma Help Center description: '# Introduction Welcome to Avoma''s developer hub.' title: Avoma Transcriptions API version: 1.0.0 servers: - description: Avoma APIs url: https://api.avoma.com security: - bearerAuth: [] tags: - description: Transcription is generated for meeting's recording. This is available once the meeting recording is processed. Availability of transcription can be verified using `transcript_ready` parameter in get meeting response. name: Transcriptions paths: /v1/transcriptions/: get: description: Returns list of transcriptions. If meeting_uuid is provided, returns transcription for that meeting directly. Otherwise returns paginated response of all transcriptions. parameters: - description: To retrieve transcriptions for meetings started at or started after given date-time(UTC). Not required when `meeting_uuid' is provided in: query name: from_date required: true schema: format: date-time type: string - description: To retrieve transcriptions for meetings started at or started before given date-time(UTC). Not required when `meeting_uuid' is provided in: query name: to_date required: true schema: format: date-time type: string - description: Unique ID of the meeting for which transcriptions will be fetched. in: query name: meeting_uuid schema: format: uuid type: string - description: Page number for pagination (only used when meeting_uuid is not provided) in: query name: page schema: minimum: 1 type: integer - description: Number of items per page (max 100, only used when meeting_uuid is not provided) in: query name: page_size schema: maximum: 100 minimum: 1 type: integer - description: Comma-separated list of attendee emails to filter meetings by.
Meetings with ***any*** of the attendees will be returned. in: query name: attendee_emails schema: example: john.doe@example.com,kelly@acme.com type: string - description: Comma-separated list of CRM account external IDs to filter by. in: query name: crm_account_ids schema: type: string - description: Comma-separated list of CRM opportunity external IDs to filter by. in: query name: crm_opportunity_ids schema: type: string - description: Comma-separated list of CRM contact external IDs to filter by. in: query name: crm_contact_ids schema: type: string - description: Comma-separated list of CRM lead external IDs to filter by. in: query name: crm_lead_ids schema: type: string responses: '200': content: application/json: schema: oneOf: - description: Single transcription response when meeting_uuid is provided example: - meeting_uuid: 123e4567-e89b-12d3-a456-426614174000 speakers: - email: user@example.com id: 0 is_rep: true name: FirstName LastName transcript: - speaker_id: 0 timestamps: - 0.77 - 0.89 - 1.04 - 1.19 transcript: Hey, how are you? transcription_vtt_url: string uuid: 095be615-a8ad-4c33-8e9c-c7612fbf6c9f items: description: Transcription object properties: meeting_uuid: description: Unique ID of the meeting that this transcription belongs to. format: uuid type: string speakers: items: properties: email: description: Email of the speaker. format: email type: string id: description: ID of the speaker. type: integer is_rep: description: '- `true`: Speaker is a representative (any speaker from the organization) - `false`: Speaker is a customer/prospect. ' type: boolean name: description: Name of the speaker. example: FirstName LastName type: - string - 'null' required: - email - id - is_rep type: object type: array transcript: example: - speaker_id: 0 timestamps: - 0.77 - 0.89 - 1.04 - 1.19 transcript: Hey, how are you? items: properties: speaker_id: description: 'Id of the speaker whom transcript paragraph belongs. Use Speakers Array to get more information like email/name based on the ''id'' of the speaker. ' type: integer timestamps: description: 'List of timestamps(in seconds) for each word in the `transcript` field. These timestamps are relative to actual (not scheduled) start of the meeting. ' example: - 0.77 - 0.89 - 1.04 - 1.19 items: format: float type: number type: array transcript: description: Actual textual transcript. example: Hey, how are you? type: string required: - transcript - timestamps - speaker_id type: object type: array transcription_vtt_url: description: Signed url of the transcription vtt file. format: url type: string uuid: description: Unique ID of the transcription. format: uuid type: string required: - uuid - transcript - speakers - transcription_vtt_url - meeting_uuid type: object title: Single Transcription type: array - description: Paginated response when meeting_uuid is not provided properties: count: description: Total number of items type: integer next: description: URL for the next page of results type: - string - 'null' previous: description: URL for the previous page of results type: - string - 'null' results: description: Array of transcription objects items: description: Transcription object properties: meeting_uuid: description: Unique ID of the meeting that this transcription belongs to. format: uuid type: string speakers: items: properties: email: description: Email of the speaker. format: email type: string id: description: ID of the speaker. type: integer is_rep: description: '- `true`: Speaker is a representative (any speaker from the organization) - `false`: Speaker is a customer/prospect. ' type: boolean name: description: Name of the speaker. example: FirstName LastName type: - string - 'null' required: - email - id - is_rep type: object type: array transcript: example: - speaker_id: 0 timestamps: - 0.77 - 0.89 - 1.04 - 1.19 transcript: Hey, how are you? items: properties: speaker_id: description: 'Id of the speaker whom transcript paragraph belongs. Use Speakers Array to get more information like email/name based on the ''id'' of the speaker. ' type: integer timestamps: description: 'List of timestamps(in seconds) for each word in the `transcript` field. These timestamps are relative to actual (not scheduled) start of the meeting. ' example: - 0.77 - 0.89 - 1.04 - 1.19 items: format: float type: number type: array transcript: description: Actual textual transcript. example: Hey, how are you? type: string required: - transcript - timestamps - speaker_id type: object type: array transcription_vtt_url: description: Signed url of the transcription vtt file. format: url type: string uuid: description: Unique ID of the transcription. format: uuid type: string required: - uuid - transcript - speakers - transcription_vtt_url - meeting_uuid type: object type: array title: Paginated Transcriptions type: object description: OK, returns transcription object when meeting_uuid is provided, otherwise returns paginated response. summary: List Transcriptions tags: - Transcriptions operationId: getV1Transcriptions x-operation-id-source: derived /v1/transcriptions/{uuid}/: get: description: Returns a single transcription for a given `uuid`. parameters: - in: path name: uuid required: true schema: description: Unique ID of the transcription. format: uuid type: string responses: '200': content: application/json: schema: description: Transcription object properties: meeting_uuid: description: Unique ID of the meeting that this transcription belongs to. format: uuid type: string speakers: items: properties: email: description: Email of the speaker. format: email type: string id: description: ID of the speaker. type: integer is_rep: description: '- `true`: Speaker is a representative (any speaker from the organization) - `false`: Speaker is a customer/prospect. ' type: boolean name: description: Name of the speaker. example: FirstName LastName type: - string - 'null' required: - email - id - is_rep type: object type: array transcript: example: - speaker_id: 0 timestamps: - 0.77 - 0.89 - 1.04 - 1.19 transcript: Hey, how are you? items: properties: speaker_id: description: 'Id of the speaker whom transcript paragraph belongs. Use Speakers Array to get more information like email/name based on the ''id'' of the speaker. ' type: integer timestamps: description: 'List of timestamps(in seconds) for each word in the `transcript` field. These timestamps are relative to actual (not scheduled) start of the meeting. ' example: - 0.77 - 0.89 - 1.04 - 1.19 items: format: float type: number type: array transcript: description: Actual textual transcript. example: Hey, how are you? type: string required: - transcript - timestamps - speaker_id type: object type: array transcription_vtt_url: description: Signed url of the transcription vtt file. format: url type: string uuid: description: Unique ID of the transcription. format: uuid type: string required: - uuid - transcript - speakers - transcription_vtt_url - meeting_uuid type: object description: OK '404': content: application/json: schema: properties: detail: description: Error details example: Not Found type: string required: - detail type: object description: Not Found summary: Get Transcription tags: - Transcriptions operationId: getV1TranscriptionsByUuid x-operation-id-source: derived components: securitySchemes: bearerAuth: description: Avoma uses bearer token for authentication of the client. For now, you will have to contact [Avoma Support](mailto:help@avoma.com) to get one issued for you. scheme: bearer type: http x-provenance: method: harvested publisher: Avoma source: https://dev.avoma.com/openapi.yml note: split by refine-openapis from the provider's own published document, verified by fetch (openapi/_original/*.fetch.json); stamped by stamp-openapi-provenance.py (roadmap#793)