> ## Documentation Index > Fetch the complete documentation index at: https://edenai-auto-update-feature-docs.mintlify.site/llms.txt > Use this file to discover all available pages before exploring further. # Audio Speech > OpenAI-compatible text-to-speech endpoint. Accepts a JSON body (``model``, ``input``, ``voice`` plus optional ``response_format`` / ``speed`` / ``instructions``) and returns raw audio bytes; ``cost`` and ``provider`` are returned in the ``x-edenai-*`` response headers. ## OpenAPI ````yaml https://api.edenai.run/v3/docs/openapi.json post /v3/audio/speech openapi: 3.1.0 info: title: Eden AI API V3 version: 3.0.0 servers: - url: https://api.edenai.run description: Production server security: [] paths: /v3/audio/speech: post: tags: - Audio summary: Audio Speech description: >- OpenAI-compatible text-to-speech endpoint. Accepts a JSON body (``model``, ``input``, ``voice`` plus optional ``response_format`` / ``speed`` / ``instructions``) and returns raw audio bytes; ``cost`` and ``provider`` are returned in the ``x-edenai-*`` response headers. operationId: audio_speech_v3_audio_speech_post requestBody: content: application/json: schema: $ref: '#/components/schemas/SpeechBody' required: true responses: '200': description: Successful Response content: application/json: schema: {} audio/mpeg: schema: type: string format: binary audio/ogg: schema: type: string format: binary audio/aac: schema: type: string format: binary audio/flac: schema: type: string format: binary audio/wav: schema: type: string format: binary audio/pcm: schema: type: string format: binary '422': description: Validation Error content: application/json: schema: $ref: '#/components/schemas/HTTPValidationError' security: - AuthBearer: [] components: schemas: SpeechBody: properties: routing: anyOf: - $ref: '#/components/schemas/ProviderRoutingPreferences' - type: 'null' description: >- How to pick between the providers that serve the requested model. Applies when `model` is a model name with no provider prefix (e.g. 'gpt-5.5'); ignored for a concrete 'provider/model' id, which already names its provider. With model='@edenai' the platform chooses the model too: `quality_cost` steers that choice, and the provider fields apply whenever the chosen model is a provider-less name. model: type: string title: Model description: provider/model, e.g. 'openai/tts-1' input: type: string title: Input description: The text to synthesize into audio. voice: type: string title: Voice description: Voice preset, e.g. 'alloy'. response_format: anyOf: - type: string - type: 'null' title: Response Format description: >- Audio format: 'mp3', 'opus', 'aac', 'flac', 'wav', or 'pcm'. Defaults to 'mp3'. Note: Gemini TTS models always return WAV and ignore this field. speed: anyOf: - type: number - type: 'null' title: Speed description: >- Playback speed. OpenAI/Azure TTS accept 0.25-4.0; other providers may use a different range or ignore it. instructions: anyOf: - type: string - type: 'null' title: Instructions description: Optional guidance for voice and delivery style. type: object required: - model - input - voice title: SpeechBody description: |- OpenAI-compatible text-to-speech request. Synthesizes ``input`` text into audio with the given ``voice``. Unknown top-level fields are dropped. HTTPValidationError: properties: detail: items: $ref: '#/components/schemas/ValidationError' type: array title: Detail type: object title: HTTPValidationError ProviderRoutingPreferences: properties: sort: anyOf: - type: string enum: - cost - speed - latency - exact - type: 'null' title: Sort description: >- What to optimise for when several providers serve the requested model. 'cost' (default) picks the cheapest for this request's shape; 'speed' the highest tokens/second; 'latency' the fastest to first token; 'exact' the most reliable at producing well-formed tool calls / structured output. Health is always a filter first — no mode will route you to a failing provider. Can also be written as a model suffix, e.g. 'gpt-5.5:speed'. sticky: anyOf: - type: boolean - type: 'null' title: Sticky description: >- Keep a conversation on the provider holding its prompt cache. On by default, and only ever active for models whose providers discount cache reads. Set false to route every request independently on price instead. Naming an explicit `sort` also takes priority over cache affinity. allow_fallbacks: type: boolean title: Allow Fallbacks description: >- Whether other providers of the same model may be tried when the chosen one fails. Set false to pin the request to the single best provider: it then fails rather than silently moving to another seller. useful when a cache-warm prompt would cold-miss elsewhere. This governs PROVIDERS of the requested model only; models you list in `fallbacks` are your own choice and are always kept. default: true quality_cost: anyOf: - type: integer maximum: 10 minimum: 0 - type: 'null' title: Quality Cost description: >- Only with model='@edenai': how far to trade answer quality for cost when the platform chooses the MODEL. 0 asks for the best model for the request, 10 for the cheapest model that can still handle it, values in between blend the two; omit it to leave the choice to the platform (quality first). This is the one `routing` field that steers the model rather than the provider — `sort` never changes which model is chosen. allowed_providers: anyOf: - items: type: string type: array - type: 'null' title: Allowed Providers description: >- Restrict routing to these providers, e.g. ['openai', 'anthropic']. Only providers that serve the requested model are considered, so an entry that does not sell it is simply inert. If none of them do, the request fails rather than falling back to a provider you excluded. Case-insensitive. Applies to routed providers only. a concrete 'provider/model' you named in `fallbacks` is your own choice and is kept. type: object title: ProviderRoutingPreferences description: >- How to choose between SELLERS of one model — and, with ``@edenai``, how far to trade quality for cost when the platform chooses the model. The seller fields are only meaningful when `model` is a canonical name (`gpt-5.5`) rather than a concrete `provider/model` — with a concrete id there is nothing to choose between. For choosing the MODEL itself see ``router_candidates`` and ``model="@edenai"``, which is a different router; ``quality_cost`` below is the one field here that speaks to it. ValidationError: properties: loc: items: anyOf: - type: string - type: integer type: array title: Location msg: type: string title: Message type: type: string title: Error Type type: object required: - loc - msg - type title: ValidationError securitySchemes: AuthBearer: type: http scheme: bearer ````