> ## Documentation Index > Fetch the complete documentation index at: https://docs.deepshi.ai/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > The Deepshi API is an OpenAI-compatible gateway. Base URL: https://api.deepshi.ai/v1. API keys start with sk-bf- and go in the Authorization: Bearer header. Prefer the official OpenAI SDKs pointed at the Deepshi base URL. Model ids are clean with no provider prefix (e.g. deepshi-3.0, claude-opus-4.8, gpt-5.5). Chat and image requests are synchronous; video and music requests are asynchronous job APIs (create, then poll). Every synchronous response carries usage.cost.total_cost in USD. Do not reference the internal /api/* admin plane or virtual keys. # Create a music job > Create an asynchronous music generation job from a style prompt. Returns a job with `status: "in_progress"`; poll `GET /v1/audio/generations/{audio_id}` until it is `completed`. Some models take lyrics and generate vocal songs, others are instrumental; call `GET /v1/models` to see the parameters a model supports. Music is billed once, on completion; pricing is shown on deepshi.ai. ## OpenAPI ````yaml /api-reference/openapi.json post /v1/audio/generations openapi: 3.1.0 info: title: Deepshi API description: >- OpenAI-compatible REST API for Deepshi's own models and leading third-party models. Authenticate with your Deepshi API key as a bearer token, and reference a model by its id (e.g. `deepshi-3.0`, `gpt-4o`) in the `model` field. version: 1.0.0 servers: - url: https://api.deepshi.ai description: Deepshi API security: - BearerAuth: [] paths: /v1/audio/generations: post: tags: - Music summary: Create a music job description: >- Create an asynchronous music generation job from a style prompt. Returns a job with `status: "in_progress"`; poll `GET /v1/audio/generations/{audio_id}` until it is `completed`. Some models take lyrics and generate vocal songs, others are instrumental; call `GET /v1/models` to see the parameters a model supports. Music is billed once, on completion; pricing is shown on deepshi.ai. operationId: createMusic requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/AudioGenerationRequest' example: model: minimax-music-v2 prompt: a warm lofi hip hop beat with mellow piano and vinyl crackle lyrics_prompt: |- [Verse] city lights fade slow in the quiet night we go [Chorus] hold on to the glow responses: '200': description: The created music job. content: application/json: schema: $ref: '#/components/schemas/AudioJob' example: id: aud_a1b2c3 object: audio model: minimax-music-v2 status: in_progress '400': $ref: '#/components/responses/BadRequest' '401': $ref: '#/components/responses/Unauthorized' '402': $ref: '#/components/responses/InsufficientQuota' '403': $ref: '#/components/responses/Forbidden' '429': $ref: '#/components/responses/RateLimited' '500': $ref: '#/components/responses/InternalError' components: schemas: AudioGenerationRequest: type: object description: >- Music generation request. `model` and `prompt` are required; `lyrics_prompt` is required by some models. The remaining fields are optional, model-specific options; call `GET /v1/models` for the parameters each model accepts. Additional parameters not listed here may also be sent and are forwarded to the model. required: - model - prompt additionalProperties: true properties: model: type: string description: The music model id. See `GET /v1/models`. example: minimax-music-v2 prompt: type: string description: >- The music style: genre, mood, instruments, tempo. Up to the model's `prompt_max_length`. example: a warm lofi hip hop beat with mellow piano and vinyl crackle lyrics_prompt: type: string description: >- The lyrics to sing, with `[Verse]` / `[Chorus]` structure tags on their own lines (separate lines with `\n`). 10 to 3000 characters. Required by some models. example: |- [Verse] city lights fade slow in the quiet night we go [Chorus] hold on to the glow is_instrumental: type: boolean description: Generate a vocal-free track, where supported. music_length_ms: type: integer description: >- Output length in milliseconds, where the model supports duration control. seconds_total: type: integer description: Output length in seconds, where the model supports duration control. duration: type: integer description: Output length in seconds, where the model supports duration control. sample_rate: type: integer description: >- Output sample rate in Hz; one of the model's allowed values, where supported. example: 44100 bitrate: type: integer description: >- Output bitrate in bits per second; one of the model's allowed values, where supported. example: 256000 format: type: string description: Output container, where supported. example: mp3 AudioJob: type: object description: A music generation job. properties: id: type: string description: The job id. Pass it to the retrieve and content endpoints. object: type: string description: Object type, always `audio`. example: audio model: type: string description: The music model id you requested. status: type: string enum: - in_progress - completed - failed description: Lifecycle state of the job. seconds: type: string description: >- Length of the generated track in seconds, present when `status` is `completed`. created_at: type: integer format: int64 description: Unix timestamp when the job was created. completed_at: type: integer format: int64 nullable: true description: Unix timestamp when the job completed. audio: type: array items: $ref: '#/components/schemas/AudioOutput' description: The generated song, present when `status` is `completed`. error: type: object nullable: true description: >- Present when `status` is `failed`. Explains the failure; the job is not charged. properties: code: type: string message: type: string AudioOutput: type: object properties: type: type: string description: Output kind, always `url`. example: url url: type: string description: Direct link to the finished `.mp3`. content_type: type: string description: Media type, always `audio/mpeg`. Error: type: object description: Error envelope, matching the OpenAI shape. required: - error properties: error: type: object properties: message: type: string description: Human-readable description of the error. type: type: string description: >- Error category, e.g. `invalid_request_error` or `insufficient_quota`. code: type: string nullable: true description: >- Machine-readable error code, e.g. `invalid_api_key`, `model_not_found`, `insufficient_quota` (may be null). param: type: string nullable: true description: The request parameter the error relates to, when applicable. responses: BadRequest: description: Malformed request, or a model id that doesn't exist in the catalog. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: The model does not exist or you do not have access to it. type: invalid_request_error code: model_not_found param: null Unauthorized: description: Missing, malformed, or unknown API key. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: Incorrect API key provided. type: invalid_request_error code: invalid_api_key param: null InsufficientQuota: description: The account is out of credits. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: You have insufficient credits to complete this request. type: insufficient_quota code: insufficient_quota param: null Forbidden: description: The key is revoked, or not permitted to use the requested model. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: The model does not exist or you do not have access to it. type: invalid_request_error code: model_not_found param: null RateLimited: description: Too many requests. Back off and retry. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: Rate limit reached. Please retry after a short delay. type: rate_limit_error code: rate_limit_exceeded param: null InternalError: description: Temporary server error. Retry with backoff. content: application/json: schema: $ref: '#/components/schemas/Error' example: error: message: >- The server had an error while processing your request. Please retry. type: api_error code: null param: null securitySchemes: BearerAuth: type: http scheme: bearer description: 'Your Deepshi API key, sent as `Authorization: Bearer `.' ````