> ## Documentation Index > Fetch the complete documentation index at: https://docs.opper.ai/llms.txt > Use this file to discover all available pages before exploring further. # Generate audio > Generate audio from a text prompt with any audio generation model from `GET /v3/audio/models`: music (`type=music`) or sound effects and ambience (`type=sound`). By default runs synchronously and returns it inline as base64 (200). Set `async: true` to run it on the background worker and get a 202 with a status URL to poll instead; use it for long tracks, which can take minutes. `model` is required, and `prompt` too unless `sections` are sent; `duration_seconds`, `instrumental`, `format` (omit it for the model's native format) and `seed` are normalized, and a value a model cannot honor answers 400 naming the limit. `sections` (a timed plan: each section's `duration_ms`, `prompt` and `styles`) needs a model with the `audio_sections` capability, and `loop` one with `audio_loop`; a model without the capability answers 400 rather than ignoring the field. Everything in `parameters` is forwarded verbatim to the provider. Lyrics are returned when the provider produces them. Set `store: true` to also save the output to /v3/files and get a reusable `file_id`; nothing is stored otherwise. Music, songs, sound effects and ambience. See the [Audio guide](/build/multimodal/audio#music-and-sound-effects) for runnable examples, timed sections and loops. ## OpenAPI ````yaml post /v3/audio/generations openapi: 3.1.0 info: description: Schema-driven generative API that orchestrates LLM-powered workflows. title: Task API version: 3.0.0 servers: - description: Production url: https://api.opper.ai - description: Local development url: http://localhost:8080 security: - BearerAuth: [] tags: - description: Schema-driven function management and execution name: Functions - description: OpenAI-compatible chat completions name: Chat - description: OpenAI Responses API compatible endpoint name: Responses - description: Google-compatible interactions endpoint name: Interactions - description: Model registry and capabilities name: Models - description: Synchronous image generation name: Images - description: Text-to-speech and speech-to-text name: Audio - description: Asynchronous video generation name: Videos - description: Reusable file storage for media inputs and generated outputs name: Files - description: Async generation status and downloads name: Artifacts - description: OpenAI-compatible embeddings name: Embeddings - description: Recorded HTTP request/response generations name: Generations - description: System health and status name: System - description: Roundtable endpoint — fan out a query to multiple LLMs and combine results name: Roundtable - description: Web search, fetch, and other utility tools name: Tools - description: Caller identity, credits, and usage name: Account - description: >- Programmatic project and API-key management. Authenticates with an `op-mak-…` management token; available on the control_plane and enterprise plans. name: Management paths: /v3/audio/generations: post: tags: - Audio summary: Generate audio description: >- Generate audio from a text prompt with any audio generation model from `GET /v3/audio/models`: music (`type=music`) or sound effects and ambience (`type=sound`). By default runs synchronously and returns it inline as base64 (200). Set `async: true` to run it on the background worker and get a 202 with a status URL to poll instead; use it for long tracks, which can take minutes. `model` is required, and `prompt` too unless `sections` are sent; `duration_seconds`, `instrumental`, `format` (omit it for the model's native format) and `seed` are normalized, and a value a model cannot honor answers 400 naming the limit. `sections` (a timed plan: each section's `duration_ms`, `prompt` and `styles`) needs a model with the `audio_sections` capability, and `loop` one with `audio_loop`; a model without the capability answers 400 rather than ignoring the field. Everything in `parameters` is forwarded verbatim to the provider. Lyrics are returned when the provider produces them. Set `store: true` to also save the output to /v3/files and get a reusable `file_id`; nothing is stored otherwise. operationId: createAudioGeneration requestBody: content: application/json: schema: $ref: '#/components/schemas/AudioGenerationRequest' required: true responses: '200': content: application/json: schema: $ref: '#/components/schemas/AudioGenerationResponse' description: Successful response '202': content: application/json: schema: $ref: '#/components/schemas/AudioGenerationJobResponse' description: >- Accepted: async generation submitted (async: true). Poll the returned `status_url` (GET /v3/artifacts/{id}/status) for a presigned download URL when complete. '400': content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' description: Bad request '401': description: Unauthorized - missing or invalid API key '500': content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' description: Internal server error components: schemas: AudioGenerationRequest: properties: async: description: >- Run generation on the background worker and return a job to poll (202) instead of waiting inline (200). Use for full-length tracks and slow models. type: boolean duration_seconds: description: >- Requested length in seconds (up to 600). Omit it to let the model choose. Limits differ by model, and some models have no length setting; a value a model cannot honor answers 400 naming its limit. type: number format: description: >- Output audio format: mp3, wav, opus or pcm. Omit it to get the model's native format. A format the model cannot produce answers 400 naming the ones it can. type: string instrumental: description: 'Music only: generate without vocals.' type: boolean loop: description: >- Generate audio that repeats seamlessly, for ambience and hums. Needs a model with the audio_loop capability (GET /v3/audio/models?capability=audio_loop); other models answer 400. type: boolean model: description: >- Audio generation model id. List them with GET /v3/audio/models: ?type=music for songs and scores, ?type=sound for sound effects and ambience, ?capability=audio_sections or ?capability=audio_loop for models that take those fields. For example elevenlabs/music_v2_5 (music) or elevenlabs/eleven_text_to_sound_v2 (sound effects). type: string parameters: description: >- Provider-specific fields, forwarded verbatim and never gated: the provider validates them. Use it for settings specific to one model, in that provider's own field names. type: object prompt: description: >- What to generate. For music: genre, mood, instruments, lyrics. For a sound effect: the sound itself, for example "a single hammer strike on an iron anvil". Required unless sections are set. type: string sections: description: >- A timed plan for music that must fit a picture: each section's duration_ms plus its own prompt and styles, in order. The sections set the total length, so duration_seconds is not sent with them. Needs a model with the audio_sections capability (GET /v3/audio/models?capability=audio_sections); other models answer 400. items: properties: duration_ms: description: >- Length of this section in milliseconds. Models set their own per-section limits and answer 400 outside them. type: integer lyrics: description: >- Words to be sung in this section. Leave it empty for an instrumental section. type: string negative_styles: description: Styles this section should avoid, for example ["vocals"]. items: type: string type: array prompt: description: >- What this section should sound like, in words, for example "a quiet piano intro" or "strings build to a swell". It describes the music and is never sung; put sung words in lyrics. type: string styles: description: >- Styles this section should have, for example ["sparse piano", "instrumental"]. items: type: string type: array required: - duration_ms type: object type: array seed: description: >- Seed for more repeatable output. Models without seed support answer 400. type: integer store: description: >- Persist the generated audio to /v3/files and return a reusable file_id. Defaults to false; set true to store. type: boolean required: - model type: object AudioGenerationResponse: properties: audio: properties: b64_json: type: string file_id: type: string mime_type: type: string url: type: string type: object created: type: integer id: type: string lyrics: type: string model: type: string usage: properties: cost: type: number duration_seconds: type: number required: - cost type: object required: - id - model - created - audio - usage type: object AudioGenerationJobResponse: properties: id: type: string status_url: type: string required: - id - status_url type: object ErrorResponse: properties: error: properties: code: type: string details: description: Any value message: type: string required: - code - message type: object meta: type: object required: - error type: object securitySchemes: BearerAuth: bearerFormat: API Key description: API key authentication. Pass your API key as a Bearer token. scheme: bearer type: http ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.