Recursal AI, Inc. Chat API

openapi: 3.1.0
info:
  title: Featherless AI Serverless Inference Account Chat API
  version: '1.0'
  description: Faithful capture of the publicly documented, OpenAI-compatible serverless inference API operated by Recursal AI, Inc. under the Featherless AI brand. Only endpoints, methods, the Bearer security scheme, and parameters that Featherless publishes in its developer documentation are represented here; request/response schemas are intentionally minimal and honest rather than exhaustive. Base URL and auth per https://featherless.ai/docs/quickstart-guide.
  x-generated: '2026-07-21'
  x-method: generated
  x-source: https://featherless.ai/docs/api-overview-and-common-options
  x-note: Featherless does not publish a machine-readable OpenAPI document (api.featherless.ai/openapi.json -> 404); this is an API Evangelist capture of the documented REST surface, not the provider's own specification.
servers:
- url: https://api.featherless.ai/v1
  description: Featherless AI production inference endpoint (OpenAI-compatible)
security:
- bearerAuth: []
tags:
- name: Chat
  description: OpenAI-compatible chat and text completion generation
paths:
  /chat/completions:
    post:
      tags:
      - Chat
      operationId: createChatCompletion
      summary: Create a chat completion
      description: Generate a model response for a chat conversation. OpenAI-compatible; set stream=true for server-sent-event token streaming.
      parameters:
      - $ref: '#/components/parameters/HttpReferer'
      - $ref: '#/components/parameters/XTitle'
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ChatCompletionRequest'
      responses:
        '200':
          description: Chat completion response
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ChatCompletionResponse'
            text/event-stream:
              schema:
                type: string
                description: Server-sent-event stream of completion chunks
        '400':
          $ref: '#/components/responses/ColdModel'
        '401':
          $ref: '#/components/responses/Unauthenticated'
        '403':
          $ref: '#/components/responses/GatedModel'
        '429':
          $ref: '#/components/responses/ConcurrencyExceeded'
        '500':
          $ref: '#/components/responses/InternalError'
        '503':
          $ref: '#/components/responses/CapacityUnavailable'
  /completions:
    post:
      tags:
      - Chat
      operationId: createCompletion
      summary: Create a text completion
      description: OpenAI-compatible text (prompt) completion.
      parameters:
      - $ref: '#/components/parameters/HttpReferer'
      - $ref: '#/components/parameters/XTitle'
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CompletionRequest'
      responses:
        '200':
          description: Text completion response
          content:
            application/json:
              schema:
                type: object
        '400':
          $ref: '#/components/responses/ColdModel'
        '401':
          $ref: '#/components/responses/Unauthenticated'
        '403':
          $ref: '#/components/responses/GatedModel'
        '429':
          $ref: '#/components/responses/ConcurrencyExceeded'
        '500':
          $ref: '#/components/responses/InternalError'
        '503':
          $ref: '#/components/responses/CapacityUnavailable'
  /tokenize:
    post:
      tags:
      - Chat
      operationId: tokenize
      summary: Tokenize text
      description: Return the tokenization of the supplied text for a given model.
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              properties:
                model:
                  type: string
                text:
                  type: string
              required:
              - model
              - text
      responses:
        '200':
          description: Tokenization result
          content:
            application/json:
              schema:
                type: object
        '401':
          $ref: '#/components/responses/Unauthenticated'
components:
  responses:
    InternalError:
      description: Internal server error on Featherless infrastructure.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    CapacityUnavailable:
      description: Service temporarily unavailable due to insufficient model capacity.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    ConcurrencyExceeded:
      description: The request would exceed the plan's reserved concurrent-unit budget.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    ColdModel:
      description: The model is cold / not yet loaded into GPU for inference.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    Unauthenticated:
      description: API key is invalid or unrecognized.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    GatedModel:
      description: The model is gated; the account must accept the model license first.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  schemas:
    CompletionRequest:
      type: object
      required:
      - model
      - prompt
      properties:
        model:
          type: string
        prompt:
          type: string
        stream:
          type: boolean
        max_tokens:
          type: integer
        temperature:
          type: number
    ChatMessage:
      type: object
      required:
      - role
      - content
      properties:
        role:
          type: string
          enum:
          - system
          - user
          - assistant
        content:
          type: string
    ChatCompletionRequest:
      type: object
      required:
      - model
      - messages
      properties:
        model:
          type: string
          description: Hugging Face model id, e.g. Qwen/Qwen2.5-7B-Instruct.
        messages:
          type: array
          items:
            $ref: '#/components/schemas/ChatMessage'
        stream:
          type: boolean
          description: Stream the response as server-sent events.
        max_tokens:
          type: integer
        temperature:
          type: number
        top_p:
          type: number
    ChatCompletionResponse:
      type: object
      properties:
        id:
          type: string
        object:
          type: string
        model:
          type: string
        choices:
          type: array
          items:
            type: object
    Error:
      type: object
      properties:
        error:
          type: object
          properties:
            message:
              type: string
            type:
              type: string
            code:
              type: string
  parameters:
    HttpReferer:
      name: HTTP-Referer
      in: header
      required: false
      description: Optional header identifying the calling application/site.
      schema:
        type: string
    XTitle:
      name: X-Title
      in: header
      required: false
      description: Optional header naming the calling application for support attribution.
      schema:
        type: string
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      description: 'Featherless API key supplied as an HTTP Bearer token (Authorization: Bearer FEATHERLESS_API_KEY). Keys are created in account settings.'