> ## Documentation Index > Fetch the complete documentation index at: https://docs.prisme.ai/llms.txt > Use this file to discover all available pages before exploring further. # Transcribe audio into text > OpenAI-compatible speech-to-text endpoint. An official OpenAI SDK pointed at the gateway calls it unchanged (`client.audio.transcriptions.create(...)`): set the SDK base URL to `https://{host}/v2/workspaces/slug:llm-gateway/webhooks/v1` and pass a user access token as the API key. This is the endpoint behind chat dictation. The audio is sent as a multipart `file` part. It is processed in memory and never stored: it is not written to storage, to the activity feed, or to logs. **Model resolution.** When `model` is omitted, the gateway uses the organization's default transcription model, then the platform default transcription model. A model not declared on any provider is answered with `400`. **Response formats.** `json` (default) and `verbose_json` return JSON, `text` returns `text/plain`. `srt` and `vtt` are refused with `400 UNSUPPORTED_RESPONSE_FORMAT`. **Permission.** Requires `llm-gateway:models:audio-transcriptions`, granted by default to the system Member role. **Governance.** The organization's allowed models apply: a model outside the allowlist is refused with `403 MODEL_NOT_ALLOWED`, or, under a `soft_downgrade` quota policy, replaced by the default transcription model. **Analytics.** Each successful call emits an `analytics.llm.transcription` event, so usage appears in analytics. ## OpenAPI ````yaml /api-reference/llm-gateway/swagger.yml post /v1/audio/transcriptions openapi: 3.0.3 info: version: 1.0.0 title: LLM Gateway API description: | Public REST API for the Prisme.ai LLM Gateway - OpenAI-compatible chat completions and embeddings, plus a managed model catalogue with governance overrides per organization. The gateway abstracts multi-provider LLM access (OpenAI, Azure OpenAI, Anthropic, Google Vertex, AWS Bedrock, OpenAI-compatible providers) behind an OpenAI-compatible request/response shape. It enforces per-tenant governance (allowed models, default models, quotas) and emits analytics events (`analytics.llm.completion`) usable for cost and carbon reporting. This spec documents only the public REST surface (endpoints exposed via Prisme.ai workspace webhooks). Internal helpers (private automations prefixed with `_`) and load-test mocks are not part of the public contract. contact: name: Prisme.ai url: https://prisme.ai servers: - url: https://{host}/v2/workspaces/slug:llm-gateway/webhooks description: Prisme.ai workspace webhooks variables: host: default: api.studio.prisme.ai description: API host (override for self-hosted or sandbox) security: - BearerAuth: [] - OrgApiKeyAuth: [] tags: - name: Completions description: OpenAI-compatible chat completions (with optional SSE streaming). - name: Embeddings description: OpenAI-compatible text embeddings. - name: Rerank description: >- Rerank documents by relevance to a query (standard rerank contract, LLM-as-reranker backend). - name: Audio description: OpenAI-compatible speech-to-text (audio transcriptions). - name: Models description: Model catalogue (CRUD + bulk replace + governance-aware listing). - name: Defaults description: >- Resolved default models for completions / embeddings / image generation / transcription / file parsing. - name: Test description: Smoke-test reachability of a model through the gateway. paths: /v1/audio/transcriptions: post: tags: - Audio summary: Transcribe audio into text description: | OpenAI-compatible speech-to-text endpoint. An official OpenAI SDK pointed at the gateway calls it unchanged (`client.audio.transcriptions.create(...)`): set the SDK base URL to `https://{host}/v2/workspaces/slug:llm-gateway/webhooks/v1` and pass a user access token as the API key. This is the endpoint behind chat dictation. The audio is sent as a multipart `file` part. It is processed in memory and never stored: it is not written to storage, to the activity feed, or to logs. **Model resolution.** When `model` is omitted, the gateway uses the organization's default transcription model, then the platform default transcription model. A model not declared on any provider is answered with `400`. **Response formats.** `json` (default) and `verbose_json` return JSON, `text` returns `text/plain`. `srt` and `vtt` are refused with `400 UNSUPPORTED_RESPONSE_FORMAT`. **Permission.** Requires `llm-gateway:models:audio-transcriptions`, granted by default to the system Member role. **Governance.** The organization's allowed models apply: a model outside the allowlist is refused with `403 MODEL_NOT_ALLOWED`, or, under a `soft_downgrade` quota policy, replaced by the default transcription model. **Analytics.** Each successful call emits an `analytics.llm.transcription` event, so usage appears in analytics. operationId: createTranscription requestBody: required: true content: multipart/form-data: schema: type: object required: - file properties: file: type: string format: binary description: > Audio file to transcribe, 25 MB maximum by default. Supported formats: `flac`, `m4a`, `mp3`, `mp4`, `mpeg`, `mpga`, `oga`, `ogg`, `wav`, `webm`. model: type: string maxLength: 256 description: | Transcription model id, for example `gpt-4o-transcribe`. Defaults to the organization's default transcription model, then to the platform default. language: type: string description: | ISO-639-1 language of the audio, for example `fr`. Improves accuracy on short recordings. Detected by the model when omitted. prompt: type: string description: >- Text that guides the transcription, such as expected vocabulary or the previous segment. temperature: type: number description: Sampling temperature, passed to the provider. response_format: type: string enum: - json - text - verbose_json default: json description: '`srt` and `vtt` are not supported.' responses: '200': description: | Transcribed text. `application/json` for `json` and `verbose_json`, `text/plain` for `text`. content: application/json: schema: type: object required: - text properties: text: type: string description: Transcribed text. usage: type: object description: Provider usage, when the provider reports it. additionalProperties: true text/plain: schema: type: string '400': description: | Invalid request: no `file` part (`FILE_REQUIRED`), `srt` or `vtt` requested (`UNSUPPORTED_RESPONSE_FORMAT`), no transcription model configured for the organization or the platform (`NO_TRANSCRIPTION_MODEL`), or a `model` declared on no provider. content: application/json: schema: $ref: '#/components/schemas/Error' '401': description: Missing or invalid authentication. content: application/json: schema: $ref: '#/components/schemas/Error' '403': description: | Missing `llm-gateway:models:audio-transcriptions` permission, or model not allowed for this caller by the organization's governance (`MODEL_NOT_ALLOWED`). content: application/json: schema: $ref: '#/components/schemas/Error' '413': description: >- Audio file larger than the size limit, 25 MB by default (`FILE_TOO_LARGE`). content: application/json: schema: $ref: '#/components/schemas/Error' '502': description: The transcription provider returned an error. content: application/json: schema: $ref: '#/components/schemas/Error' x-codeSamples: - lang: bash label: curl source: > curl https://{host}/v2/workspaces/slug:llm-gateway/webhooks/v1/audio/transcriptions \ -H "Authorization: Bearer $PRISME_TOKEN" \ -F file=@meeting-notes.m4a \ -F language=fr - lang: python label: Python (OpenAI SDK) source: | from openai import OpenAI client = OpenAI( base_url="https://{host}/v2/workspaces/slug:llm-gateway/webhooks/v1", api_key="", ) with open("meeting-notes.m4a", "rb") as audio: transcript = client.audio.transcriptions.create( model="gpt-4o-transcribe", file=audio, language="fr", ) print(transcript.text) components: schemas: Error: type: object required: - error description: | Standard error envelope. `error` carries either a stable PascalCase identifier or a free-text label (legacy endpoints) - `code` is the canonical machine-readable identifier going forward. properties: error: type: string description: Stable PascalCase identifier or short error label. message: type: string description: Human-readable error message. code: type: string description: | Machine-readable error code. Observed values include `RATE_LIMITED`, `MODEL_NOT_ALLOWED`, `HEADER_POLICY_VIOLATION`, `MODEL_NOT_FOUND`, `MODEL_EXISTS`, `MISSING_MODEL_ID`, `MISSING_TYPE`, `INVALID_BODY`, `INVALID_ITEMS`, `INVALID_DIMENSIONS`, `PAYLOAD_TOO_LARGE`, `METHOD_NOT_ALLOWED`, `PROVIDER_ERROR`, `PROVIDER_NO_RESPONSE`. details: description: Optional structured context (e.g. list of invalid items). additionalProperties: true status: type: integer description: HTTP status mirror, when present. retryAfter: type: integer description: Seconds to wait before retrying (rate-limit responses). provider: type: string description: Upstream provider name (provider-error responses). model: type: string description: Model id involved in the error (provider-error responses). provider_error_type: type: string description: Upstream provider's own error class name. securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: JWT description: | User-bound credential carrying an identity: either a session JWT or a user access token (`at:*`) generated from the user settings UI. Send as `Authorization: Bearer `. Org API keys (`iak_*`) are **not** accepted here - they carry no user identity. Use the `x-prismeai-api-key` header instead (see `OrgApiKeyAuth`). OrgApiKeyAuth: type: apiKey in: header name: x-prismeai-api-key description: | Organization API key (`iak_{orgSlug}_{uuid}`). Unlike `Authorization: Bearer`, this credential is **not** tied to a user identity - it is bound to the org and its effective access is defined by the scopes / permission rules attached to it (it can be restricted to a single project, or kept broader). Send as `x-prismeai-api-key: iak_...`. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.