first commit

docs/openapi/schemas/integrations/openai/audio.yaml

@@ -0,0 +1,90 @@
+# OpenAI Integration Audio Schemas (Speech and Transcription)
+
+# Speech (TTS) Request
+OpenAISpeechRequest:
+  type: object
+  required:
+    - model
+    - input
+  properties:
+    model:
+      type: string
+      description: Model identifier (e.g., tts-1, tts-1-hd)
+      example: tts-1
+    input:
+      type: string
+      description: Text to convert to speech
+    voice:
+      type: string
+      description: Voice to use
+      enum: [alloy, echo, fable, onyx, nova, shimmer]
+    response_format:
+      type: string
+      enum: [mp3, opus, aac, flac, wav, pcm]
+    speed:
+      type: number
+      minimum: 0.25
+      maximum: 4.0
+    stream_format:
+      type: string
+      enum: [sse]
+      description: Set to 'sse' for streaming
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+
+# Transcription Request
+OpenAITranscriptionRequest:
+  type: object
+  required:
+    - model
+    - file
+  properties:
+    model:
+      type: string
+      description: Model identifier (e.g., whisper-1)
+      example: whisper-1
+    file:
+      type: string
+      format: binary
+      description: Audio file to transcribe
+    language:
+      type: string
+      description: Language of the audio (ISO 639-1)
+    prompt:
+      type: string
+      description: Prompt to guide transcription
+    response_format:
+      type: string
+      enum: [json, text, srt, verbose_json, vtt]
+    temperature:
+      type: number
+      minimum: 0
+      maximum: 1
+    timestamp_granularities:
+      type: array
+      items:
+        type: string
+        enum: [word, segment]
+    stream:
+      type: boolean
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+
+# Responses reuse inference schemas
+OpenAISpeechResponse:
+  $ref: '../../inference/speech.yaml#/SpeechResponse'
+
+OpenAISpeechStreamResponse:
+  $ref: '../../inference/speech.yaml#/SpeechStreamResponse'
+
+OpenAITranscriptionResponse:
+  $ref: '../../inference/transcription.yaml#/TranscriptionResponse'
+
+OpenAITranscriptionStreamResponse:
+  $ref: '../../inference/transcription.yaml#/TranscriptionStreamResponse'

docs/openapi/schemas/integrations/openai/batch.yaml

@@ -0,0 +1,57 @@
+# OpenAI Integration Batch API Schemas
+# Reuses inference batch schemas since OpenAI integration uses Bifrost format
+
+# Batch Create Request - uses Bifrost format with provider field
+OpenAIBatchCreateRequest:
+  $ref: '../../inference/batch.yaml#/BatchCreateRequest'
+
+OpenAIBatchCreateResponse:
+  $ref: '../../inference/batch.yaml#/BatchCreateResponse'
+
+OpenAIBatchListRequest:
+  type: object
+  properties:
+    limit:
+      type: integer
+      description: Maximum number of batches to return
+      default: 30
+    after:
+      type: string
+      description: Cursor for pagination
+    provider:
+      type: string
+      description: Filter by provider
+      example: openai
+
+OpenAIBatchListResponse:
+  $ref: '../../inference/batch.yaml#/BatchListResponse'
+
+OpenAIBatchRetrieveRequest:
+  type: object
+  required:
+    - batch_id
+  properties:
+    batch_id:
+      type: string
+      description: Batch ID to retrieve
+    provider:
+      type: string
+      description: Provider for the batch
+
+OpenAIBatchRetrieveResponse:
+  $ref: '../../inference/batch.yaml#/BatchRetrieveResponse'
+
+OpenAIBatchCancelRequest:
+  type: object
+  required:
+    - batch_id
+  properties:
+    batch_id:
+      type: string
+      description: Batch ID to cancel
+    provider:
+      type: string
+      description: Provider for the batch
+
+OpenAIBatchCancelResponse:
+  $ref: '../../inference/batch.yaml#/BatchCancelResponse'

docs/openapi/schemas/integrations/openai/chat.yaml

@@ -0,0 +1,121 @@
+# OpenAI Integration Chat Completions Schemas
+# Reuses inference schemas where possible since Bifrost follows OpenAI format
+
+OpenAIChatRequest:
+  type: object
+  required:
+    - model
+    - messages
+  properties:
+    model:
+      type: string
+      description: Model identifier (e.g., gpt-4, gpt-3.5-turbo)
+      example: gpt-4
+    messages:
+      type: array
+      items:
+        $ref: '#/OpenAIMessage'
+      description: List of messages in the conversation
+    stream:
+      type: boolean
+      description: Whether to stream the response
+    max_tokens:
+      type: integer
+      description: Maximum tokens to generate (legacy, use max_completion_tokens)
+    max_completion_tokens:
+      type: integer
+      description: Maximum tokens to generate
+    temperature:
+      type: number
+      minimum: 0
+      maximum: 2
+    top_p:
+      type: number
+    frequency_penalty:
+      type: number
+      minimum: -2.0
+      maximum: 2.0
+    presence_penalty:
+      type: number
+      minimum: -2.0
+      maximum: 2.0
+    logit_bias:
+      type: object
+      additionalProperties:
+        type: number
+    logprobs:
+      type: boolean
+    top_logprobs:
+      type: integer
+    n:
+      type: integer
+    stop:
+      oneOf:
+        - type: string
+        - type: array
+          items:
+            type: string
+    seed:
+      type: integer
+    user:
+      type: string
+    tools:
+      type: array
+      items:
+        $ref: '../../inference/chat.yaml#/ChatTool'
+    tool_choice:
+      $ref: '../../inference/chat.yaml#/ChatToolChoice'
+    parallel_tool_calls:
+      type: boolean
+    response_format:
+      type: object
+      description: Format for the response
+    reasoning_effort:
+      type: string
+      enum: [none, minimal, low, medium, high, xhigh]
+      description: OpenAI reasoning effort level
+    service_tier:
+      type: string
+    stream_options:
+      $ref: '../../inference/chat.yaml#/ChatStreamOptions'
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+      description: Fallback models
+
+OpenAIMessage:
+  type: object
+  required:
+    - role
+  properties:
+    role:
+      type: string
+      enum: [system, user, assistant, tool, developer]
+    name:
+      type: string
+    content:
+      $ref: '../../inference/chat.yaml#/ChatMessageContent'
+    tool_call_id:
+      type: string
+      description: For tool messages
+    refusal:
+      type: string
+    reasoning:
+      type: string
+    annotations:
+      type: array
+      items:
+        $ref: '../../inference/chat.yaml#/ChatAssistantMessageAnnotation'
+    tool_calls:
+      type: array
+      items:
+        $ref: '../../inference/chat.yaml#/ChatAssistantMessageToolCall'
+
+# Response reuses inference schema since format is identical
+OpenAIChatResponse:
+  $ref: '../../inference/chat.yaml#/ChatCompletionResponse'
+
+OpenAIChatStreamResponse:
+  $ref: '../../inference/chat.yaml#/ChatCompletionStreamResponse'

docs/openapi/schemas/integrations/openai/common.yaml

@@ -0,0 +1,51 @@
+# OpenAI Integration Common Types
+
+OpenAIError:
+  type: object
+  properties:
+    error:
+      type: object
+      properties:
+        message:
+          type: string
+        type:
+          type: string
+        param:
+          type: string
+          nullable: true
+        code:
+          type: string
+          nullable: true
+
+# OpenAI uses the same model format but without provider prefix
+OpenAIModel:
+  type: object
+  properties:
+    id:
+      type: string
+      description: Model identifier
+    object:
+      type: string
+      default: model
+    owned_by:
+      type: string
+    created:
+      type: integer
+      format: int64
+    active:
+      type: boolean
+      description: GROQ-specific field
+    context_window:
+      type: integer
+      description: GROQ-specific field
+
+OpenAIListModelsResponse:
+  type: object
+  properties:
+    object:
+      type: string
+      default: list
+    data:
+      type: array
+      items:
+        $ref: '#/OpenAIModel'

docs/openapi/schemas/integrations/openai/embeddings.yaml

@@ -0,0 +1,36 @@
+# OpenAI Integration Embeddings Schemas
+
+OpenAIEmbeddingRequest:
+  type: object
+  required:
+    - model
+    - input
+  properties:
+    model:
+      type: string
+      description: Model identifier
+      example: text-embedding-3-small
+    input:
+      oneOf:
+        - type: string
+        - type: array
+          items:
+            type: string
+      description: Input text to embed
+    encoding_format:
+      type: string
+      enum: [float, base64]
+    dimensions:
+      type: integer
+      description: Number of dimensions for the embedding
+    user:
+      type: string
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+
+# Response reuses inference schema
+OpenAIEmbeddingResponse:
+  $ref: '../../inference/embeddings.yaml#/EmbeddingResponse'

docs/openapi/schemas/integrations/openai/files.yaml

@@ -0,0 +1,95 @@
+# OpenAI Integration Files API Schemas
+# Reuses inference files schemas since OpenAI integration uses Bifrost format
+
+OpenAIFileUploadRequest:
+  type: object
+  required:
+    - file
+    - purpose
+  properties:
+    file:
+      type: string
+      format: binary
+      description: File to upload
+    purpose:
+      type: string
+      enum: [assistants, assistants_output, batch, batch_output, fine-tune, fine-tune-results, vision, user_data, evals]
+      description: Purpose of the file
+    provider:
+      type: string
+      description: Provider for file storage
+    storage_config:
+      $ref: '../../inference/files.yaml#/FileStorageConfig'
+
+OpenAIFileUploadResponse:
+  $ref: '../../inference/files.yaml#/FileUploadResponse'
+
+OpenAIFileListRequest:
+  type: object
+  properties:
+    purpose:
+      type: string
+      description: Filter by purpose
+    limit:
+      type: integer
+      description: Maximum files to return
+    after:
+      type: string
+      description: Cursor for pagination
+    order:
+      type: string
+      enum: [asc, desc]
+    provider:
+      type: string
+      description: Filter by provider
+
+OpenAIFileListResponse:
+  $ref: '../../inference/files.yaml#/FileListResponse'
+
+OpenAIFileRetrieveRequest:
+  type: object
+  required:
+    - file_id
+  properties:
+    file_id:
+      type: string
+      description: File ID to retrieve
+    provider:
+      type: string
+      description: Provider for the file
+    storage_config:
+      $ref: '../../inference/files.yaml#/FileStorageConfig'
+
+OpenAIFileRetrieveResponse:
+  $ref: '../../inference/files.yaml#/FileRetrieveResponse'
+
+OpenAIFileDeleteRequest:
+  type: object
+  required:
+    - file_id
+  properties:
+    file_id:
+      type: string
+      description: File ID to delete
+    provider:
+      type: string
+      description: Provider for the file
+    storage_config:
+      $ref: '../../inference/files.yaml#/FileStorageConfig'
+
+OpenAIFileDeleteResponse:
+  $ref: '../../inference/files.yaml#/FileDeleteResponse'
+
+OpenAIFileContentRequest:
+  type: object
+  required:
+    - file_id
+  properties:
+    file_id:
+      type: string
+      description: File ID to get content for
+    provider:
+      type: string
+      description: Provider for the file
+    storage_config:
+      $ref: '../../inference/files.yaml#/FileStorageConfig'

docs/openapi/schemas/integrations/openai/images.yaml

@@ -0,0 +1,133 @@
+# OpenAI Integration - Image Generation Schemas
+
+OpenAIImageGenerationRequest:
+  type: object
+  required:
+    - model
+    - prompt
+  properties:
+    model:
+      type: string
+      description: Model identifier
+    prompt:
+      type: string
+      description: Text prompt to generate image
+    n:
+      type: integer
+      minimum: 1
+      maximum: 10
+      default: 1
+      description: Number of images to generate
+    size:
+      type: string
+      enum:
+        - "256x256"
+        - "512x512"
+        - "1024x1024"
+        - "1792x1024"
+        - "1024x1792"
+        - "1536x1024"
+        - "1024x1536"
+        - "auto"
+      description: Size of the generated image
+    quality:
+      type: string
+      enum:
+        - "standard"
+        - "hd"
+      description: Quality of the generated image
+    style:
+      type: string
+      enum:
+        - "natural"
+        - "vivid"
+      description: Style of the generated image
+    response_format:
+      type: string
+      enum:
+        - "url"
+        - "b64_json"
+      default: "url"
+      description: Format of the response. This parameter is not supported for streaming requests.
+    user:
+      type: string
+      description: User identifier for tracking
+    stream:
+      type: boolean
+      default: false
+      description: |
+        Whether to stream the response. When true, images are sent as base64 chunks via SSE.
+    fallbacks:
+      type: array
+      items:
+        type: string
+      description: Fallback models to try if primary model fails
+
+OpenAIImageGenerationResponse:
+  type: object
+  properties:
+    created:
+      type: integer
+      format: int64
+      description: Unix timestamp when the image was created
+    data:
+      type: array
+      items:
+        $ref: '../../../schemas/inference/images.yaml#/ImageData'
+      description: Array of generated images
+    background:
+      type: string
+      description: Background type used
+    output_format:
+      type: string
+      description: Output format used
+    quality:
+      type: string
+      description: Quality setting used
+    size:
+      type: string
+      description: Size setting used
+    usage:
+      $ref: '../../../schemas/inference/images.yaml#/ImageUsage'
+
+OpenAIImageStreamResponse:
+  type: object
+  description: |
+    Streaming response chunk for image generation (OpenAI format).
+    Sent via Server-Sent Events (SSE) when stream=true.
+  properties:
+    type:
+      type: string
+      enum:
+        - "image_generation.partial_image"
+        - "image_generation.completed"
+        - "error"
+      description: Type of stream event
+    b64_json:
+      type: string
+      description: Base64-encoded chunk of image data
+    partial_image_index:
+      type: integer
+      description: Index of the partial image chunk
+    sequence_number:
+      type: integer
+      description: Ordering index for stream chunks
+    created_at:
+      type: integer
+      format: int64
+      description: Timestamp when chunk was created
+    size:
+      type: string
+      description: Size of the generated image
+    quality:
+      type: string
+      description: Quality setting used
+    background:
+      type: string
+      description: Background type used
+    output_format:
+      type: string
+      description: Output format used
+    usage:
+      $ref: '../../../schemas/inference/images.yaml#/ImageUsage'
+      description: Token usage (usually in final chunk)

docs/openapi/schemas/integrations/openai/responses.yaml

@@ -0,0 +1,108 @@
+# OpenAI Integration Responses API Schemas
+
+OpenAIResponsesRequest:
+  type: object
+  required:
+    - model
+    - input
+  properties:
+    model:
+      type: string
+      description: Model identifier
+      example: gpt-4
+    input:
+      $ref: '#/OpenAIResponsesInput'
+    stream:
+      type: boolean
+    instructions:
+      type: string
+      description: System instructions for the model
+    max_output_tokens:
+      type: integer
+    metadata:
+      type: object
+      additionalProperties: true
+    parallel_tool_calls:
+      type: boolean
+    previous_response_id:
+      type: string
+    reasoning:
+      $ref: '#/OpenAIResponsesReasoning'
+    store:
+      type: boolean
+    temperature:
+      type: number
+      minimum: 0
+      maximum: 2
+    text:
+      $ref: '#/OpenAIResponsesTextConfig'
+    tool_choice:
+      $ref: '../../inference/responses.yaml#/ResponsesToolChoice'
+    tools:
+      type: array
+      items:
+        $ref: '../../inference/responses.yaml#/ResponsesTool'
+    top_p:
+      type: number
+    truncation:
+      type: string
+      enum: [auto, disabled]
+    user:
+      type: string
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+
+OpenAIResponsesInput:
+  oneOf:
+    - type: string
+    - type: array
+      items:
+        $ref: '../../inference/responses.yaml#/ResponsesMessage'
+  description: Input - can be a string or array of messages
+
+OpenAIResponsesReasoning:
+  type: object
+  properties:
+    effort:
+      type: string
+      enum: [none, minimal, low, medium, high, xhigh]
+    generate_summary:
+      type: string
+      enum: [auto, concise, detailed]
+    summary:
+      type: string
+      enum: [auto, concise, detailed]
+    max_tokens:
+      type: integer
+
+OpenAIResponsesTextConfig:
+  type: object
+  properties:
+    format:
+      $ref: '#/OpenAIResponsesTextFormat'
+
+OpenAIResponsesTextFormat:
+  type: object
+  properties:
+    type:
+      type: string
+      enum: [text, json_object, json_schema]
+    json_schema:
+      type: object
+      properties:
+        name:
+          type: string
+        schema:
+          type: object
+        strict:
+          type: boolean
+
+# Response reuses inference schema
+OpenAIResponsesResponse:
+  $ref: '../../inference/responses.yaml#/ResponsesResponse'
+
+OpenAIResponsesStreamResponse:
+  $ref: '../../inference/responses.yaml#/ResponsesStreamResponse'

docs/openapi/schemas/integrations/openai/text.yaml

@@ -0,0 +1,74 @@
+# OpenAI Integration Text Completions Schemas (Legacy Completions API)
+
+OpenAITextCompletionRequest:
+  type: object
+  required:
+    - model
+    - prompt
+  properties:
+    model:
+      type: string
+      description: Model identifier
+      example: gpt-3.5-turbo-instruct
+    prompt:
+      oneOf:
+        - type: string
+        - type: array
+          items:
+            type: string
+      description: The prompt(s) to generate completions for
+    stream:
+      type: boolean
+      description: Whether to stream the response
+    max_tokens:
+      type: integer
+    temperature:
+      type: number
+      minimum: 0
+      maximum: 2
+    top_p:
+      type: number
+    frequency_penalty:
+      type: number
+      minimum: -2.0
+      maximum: 2.0
+    presence_penalty:
+      type: number
+      minimum: -2.0
+      maximum: 2.0
+    logit_bias:
+      type: object
+      additionalProperties:
+        type: number
+    logprobs:
+      type: integer
+    n:
+      type: integer
+    stop:
+      oneOf:
+        - type: string
+        - type: array
+          items:
+            type: string
+    suffix:
+      type: string
+    echo:
+      type: boolean
+    best_of:
+      type: integer
+    user:
+      type: string
+    seed:
+      type: integer
+    # Bifrost-specific
+    fallbacks:
+      type: array
+      items:
+        type: string
+
+# Response reuses inference schema
+OpenAITextCompletionResponse:
+  $ref: '../../inference/text.yaml#/TextCompletionResponse'
+
+OpenAITextCompletionStreamResponse:
+  $ref: '../../inference/text.yaml#/TextCompletionStreamResponse'