> ## Documentation Index > Fetch the complete documentation index at: https://docs.moda.app/llms.txt > Use this file to discover all available pages before exploring further. # Generate Audio > Generate speech, music, or a sound effect from text (metered). Pick `mode` first: `text_to_speech` (a voice saying your exact words — `prompt` IS the script), `text_to_music`, or `text_to_sfx`. The model card on `GET /v1/media/models` says which modes each model serves. Synchronous: the render runs inside this call and there is no `wait: false` task lane. Speech returns in seconds. Music and sound effects can be asked for at up to 600 seconds of output and may take correspondingly longer; a render that outruns this call's wait is asked to cancel and reported as a retryable 503. Repeat that call with the same `idempotency_key`: it adopts the existing provider job, so it returns the audio if the render finished anyway and reports it cancelled if it did not — and it can never pay for a second render. Only once it reports cancelled is a new request worth making — and that one needs a NEW `idempotency_key` as well as a shorter duration or fewer samples, since a changed payload under the old key is a 409 `idempotency_conflict`. `results` carries one entry per delivered take (`num_samples`), `applied` reports the knobs that actually ran, and `adjustments` names every field that differed from the ask, with a reason — read them before describing the output to a user. ## OpenAPI ````yaml /openapi/moda-public-api.yaml post /media/generate-audio openapi: 3.1.0 info: title: Moda Public API description: > Programmatic access to Moda's canvas design platform. Create designs, export assets, manage brand kits, and run AI design tasks. ## Versioning Pin response shapes with a calendar-dated `Moda-Version` header (e.g. `2026-05-01`). **The pin is global** — it applies to every endpoint in the request, not just the one you adopted it for, so raising it to use a new endpoint also moves your other responses to that version's shapes. ### Migrating `2026-04-12` → `2026-05-01` - **Tasks** move from the flat `JobResponse` (`job_id`, `canvas_url`, `can_export`) to the canonical `Task` envelope. The export artifact is now under `result.export` as `{url, format, page_count}`; update any code that read the export URL from the old top-level fields. - **Multi-page PNG/JPEG exports are delivered as a single `.zip`** of per-page files (`page-1.png`, …), so `result.export.format` is `zip`. For one bundled document instead, request `format=pdf` or `pptx`; for a single image, pass `page_number`. version: 1.0.0 x-moda-api-version: '2026-05-01' servers: - url: https://api.moda.app/v1 description: Production security: [] tags: - name: canvases description: List, search, read, export, and share canvases - name: tasks description: Start and monitor AI design tasks - name: organizations description: List organizations and teams - name: credits description: Check credit balance and usage - name: brand-kits description: Manage brand kits - name: remix description: Duplicate and edit canvases - name: share-links description: Resolve share URLs to canvas identifiers - name: uploads description: Upload files for use as attachments - name: usage description: Aggregate API usage stats for the caller's team - name: embed description: Signed iframe embed sessions - name: web description: 'Metered web research: search and page reading' - name: websites description: >- Create, update, publish, screenshot, and manage hosted multi-page static sites - name: drive description: >- Organize the workspace: folders, the folder tree, and item move/rename/visibility/delete for folders, canvases, and files paths: /media/generate-audio: post: tags: - media summary: Generate Audio description: >- Generate speech, music, or a sound effect from text (metered). Pick `mode` first: `text_to_speech` (a voice saying your exact words — `prompt` IS the script), `text_to_music`, or `text_to_sfx`. The model card on `GET /v1/media/models` says which modes each model serves. Synchronous: the render runs inside this call and there is no `wait: false` task lane. Speech returns in seconds. Music and sound effects can be asked for at up to 600 seconds of output and may take correspondingly longer; a render that outruns this call's wait is asked to cancel and reported as a retryable 503. Repeat that call with the same `idempotency_key`: it adopts the existing provider job, so it returns the audio if the render finished anyway and reports it cancelled if it did not — and it can never pay for a second render. Only once it reports cancelled is a new request worth making — and that one needs a NEW `idempotency_key` as well as a shorter duration or fewer samples, since a changed payload under the old key is a 409 `idempotency_conflict`. `results` carries one entry per delivered take (`num_samples`), `applied` reports the knobs that actually ran, and `adjustments` names every field that differed from the ask, with a reason — read them before describing the output to a user. operationId: mediaGenerateAudio parameters: - $ref: '#/components/parameters/ModaVersion' requestBody: content: application/json: schema: $ref: '#/components/schemas/GenerateAudioRequest' required: true responses: '200': description: Successful Response content: application/json: schema: additionalProperties: true type: object title: Response Mediagenerateaudio '401': description: Authentication required. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '402': description: Insufficient credits for a metered operation. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '403': description: Permission denied for this scope. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '404': description: Resource not found. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '409': description: Conflict (idempotency / resource state). content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '422': description: Request validation failed. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '429': description: Rate limit exceeded. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' '500': description: Internal error. content: application/json: schema: $ref: '#/components/schemas/ErrorEnvelope' security: - API Key: [] components: parameters: ModaVersion: name: Moda-Version in: header required: false description: >- Calendar-dated API version pin. New integrations should pin `2026-05-01` to opt into the newest response shapes. For back-compat the server also accepts requests with no header and resolves them to the current default (today: `2026-04-12`); that default advances on each [sunset date](/api-reference/versioning). Any unsupported value returns 400 `unsupported_version`. example: '2026-05-01' schema: type: string enum: - '2026-04-12' - '2026-05-01' default: '2026-05-01' example: '2026-05-01' schemas: GenerateAudioRequest: properties: prompt: type: string minLength: 1 title: Prompt description: >- Sent to the selected model VERBATIM. For text_to_speech this IS the script — the words that get spoken, with no stage directions. For the other modes it describes the result: genre, instrumentation, mood, tempo, texture. model: type: string title: Model description: Concrete audio model id; see GET /v1/media/models. mode: type: string title: Mode description: >- text_to_speech, text_to_music, or text_to_sfx. Stated rather than derived — every mode takes exactly text, so there is no input role to derive it from. The model card on GET /v1/media/models says which modes each model serves. voice: anyOf: - type: string - type: 'null' title: Voice description: >- Speech only. Normally a preset name from the mode's voices on GET /v1/media/models; on a mode whose card sets voice_is_free_form, voices is empty and this takes any provider voice name or cloned-voice id instead, checked for shape here and resolved by the provider. Omit for the mode's default. duration_seconds: anyOf: - type: number - type: string - type: 'null' title: Duration Seconds description: >- Music and sound effects only, and it SNAPS into the mode's envelope with the adjustment reported. Speech has no duration control — a value passed there is dropped and reported rather than rejected. num_samples: anyOf: - type: integer - type: 'null' title: Num Samples description: >- Distinct takes to render, where the mode has a sample axis (max_samples on GET /v1/media/models). EVERY sample is billed, and a mode with a billing floor bills each short take at the floor. model_params: anyOf: - additionalProperties: true type: object - type: 'null' title: Model Params idempotency_key: anyOf: - type: string maxLength: 200 - type: 'null' title: Idempotency Key description: >- Keys the provider-job checkpoint: a retried call with the same key resumes the existing render instead of paying for a duplicate. additionalProperties: false type: object required: - prompt - model - mode title: GenerateAudioRequest ErrorEnvelope: properties: type: $ref: '#/components/schemas/ErrorType' code: type: string title: Code description: >- Stable machine string keyed to doc_url. Never changes once published. message: type: string title: Message description: >- Human-readable message, for developers. Not localized, not user-facing. doc_url: type: string title: Doc Url description: Permalink to the docs page for this code. request_id: anyOf: - type: string - type: 'null' title: Request Id description: Correlator for logs/Sentry/audit. causes: anyOf: - items: $ref: '#/components/schemas/ErrorEnvelope' type: array - type: 'null' title: Causes description: Aggregated upstream failures (e.g. per-page export errors). details: anyOf: - additionalProperties: true type: object - type: 'null' title: Details description: >- Code-specific detail payload (e.g. validation field list). Callers should pass ``None`` rather than ``{}`` to omit the key from the response. retry_after_ms: anyOf: - type: integer - type: 'null' title: Retry After Ms description: Hint for rate-limited or transient errors. retryable: anyOf: - type: boolean - type: 'null' title: Retryable description: >- Explicit retryability for this code, from the error-code registry: ``false`` means retrying the same request cannot succeed (fix the input / resource first); ``true`` means the condition is transient and a retry (with backoff / after the blocker clears) can succeed. Omitted for codes whose retryability is context-dependent — do not derive it from ``type`` or the HTTP status when this field is present. type: object required: - type - code - message - doc_url title: ErrorEnvelope description: >- Canonical error payload, nested under the top-level ``{"error": ...}`` key. ErrorType: type: string enum: - invalid_request - authentication - permission - not_found - conflict - rate_limited - idempotency_conflict - unprocessable - upstream_error - internal_error title: ErrorType description: Closed set of high-level categories SDKs branch on. securitySchemes: API Key: type: http description: API key from Settings > Developer > REST API scheme: bearer ````