Sahara AI Compute API

openapi: 3.1.0
info:
  title: Sahara AI Compute API
  description: >-
    OpenAI-compatible inference and model-discovery API for the Sahara AI
    Developer Platform. Exposes a routed compute layer that fans requests out
    across multiple upstream providers (OpenAI, Lepton, Together, and others),
    plus discovery endpoints for the models and providers available on the
    network. Faithful transcription of the published API documentation; not a
    provider-published OpenAPI.
  version: '1.0'
  contact:
    name: Sahara AI Developer Platform
    url: https://docs.saharaai.com/developer-docs-ai-developer-portal/api-documentation
  x-apisjson-generated: true
  x-source: https://docs.saharaai.com/developer-docs-ai-developer-portal/api-documentation.md
servers:
  - url: https://app.saharaai.com/developer/api
    description: AI Developer Platform compute API
  - url: https://portal.saharalabs.ai/api
    description: Developer Portal (Sahara Labs) compute API
security:
  - ApiKeyAuth: []
  - BearerAuth: []
tags:
  - name: Discovery
    description: Discover models and compute providers available on the network.
  - name: Inference
    description: OpenAI-compatible chat completion inference.
paths:
  /compute/models:
    get:
      operationId: listModels
      summary: List all available models
      description: >-
        Fetches all registered models across compute providers. Optionally
        filter to a single provider with the `provider` query parameter.
      tags:
        - Discovery
      parameters:
        - name: provider
          in: query
          required: false
          description: Filter models to a specific compute provider.
          schema:
            type: string
      responses:
        '200':
          description: A list of available models.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ModelList'
        '400':
          $ref: '#/components/responses/BadRequest'
        '404':
          $ref: '#/components/responses/NotFound'
        '500':
          $ref: '#/components/responses/ServerError'
  /compute/providers:
    get:
      operationId: listProviders
      summary: List all compute providers
      description: >-
        Lists all compute providers on the network (e.g. OpenAI, Lepton,
        Together). Optionally filter to the providers that serve a specific
        model with the `model` query parameter.
      tags:
        - Discovery
      parameters:
        - name: model
          in: query
          required: false
          description: Filter providers to those serving a specific model.
          schema:
            type: string
      responses:
        '200':
          description: A list of compute providers.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ProviderList'
        '400':
          $ref: '#/components/responses/BadRequest'
        '404':
          $ref: '#/components/responses/NotFound'
        '500':
          $ref: '#/components/responses/ServerError'
  /compute/modelDetail:
    get:
      operationId: getModelDetail
      summary: Get model metadata
      description: >-
        Fetch metadata and detailed usage requirements for a specific
        model-provider pair.
      tags:
        - Discovery
      parameters:
        - name: model
          in: query
          required: false
          description: Model identifier to fetch details for.
          schema:
            type: string
        - name: provider
          in: query
          required: false
          description: Provider serving the model.
          schema:
            type: string
      responses:
        '200':
          description: Model metadata and usage requirements.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ModelDetail'
        '400':
          $ref: '#/components/responses/BadRequest'
        '404':
          $ref: '#/components/responses/NotFound'
        '500':
          $ref: '#/components/responses/ServerError'
  /compute/chat/completions:
    post:
      operationId: createChatCompletion
      summary: Create a chat completion
      description: >-
        Send an OpenAI-compatible inference request. Compatible with the OpenAI
        SDK, LangChain, and the OpenAI Agents SDK. Supports streaming responses.
        The target provider may be selected via the `OpenAI-Organization`
        header.
      tags:
        - Inference
      parameters:
        - name: OpenAI-Organization
          in: header
          required: false
          description: Select the upstream compute provider to route the request to.
          schema:
            type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ChatCompletionRequest'
      responses:
        '200':
          description: An OpenAI-format chat completion.
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ChatCompletionResponse'
        '400':
          $ref: '#/components/responses/BadRequest'
        '404':
          $ref: '#/components/responses/NotFound'
        '500':
          $ref: '#/components/responses/ServerError'
components:
  securitySchemes:
    ApiKeyAuth:
      type: apiKey
      in: header
      name: x-api-key
      description: >-
        API key issued from the Developer Portal. Passed in the `x-api-key`
        header. Never expose your API key in public code or repositories.
    BearerAuth:
      type: http
      scheme: bearer
      description: >-
        API key passed as a Bearer token in the Authorization header
        (OpenAI-compatible clients).
  responses:
    BadRequest:
      description: Bad request — check request formatting.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    NotFound:
      description: Not found — verify pipeline or model IDs.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
    ServerError:
      description: Internal server error — retry or contact support.
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Error'
  schemas:
    ModelList:
      type: object
      description: A collection of models registered across providers.
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Model'
    Model:
      type: object
      properties:
        id:
          type: string
          description: Model identifier (e.g. gpt-4o).
        provider:
          type: string
          description: Compute provider serving this model.
    ProviderList:
      type: object
      description: A collection of compute providers.
      properties:
        data:
          type: array
          items:
            $ref: '#/components/schemas/Provider'
    Provider:
      type: object
      properties:
        name:
          type: string
          description: Provider name (e.g. OpenAI, Lepton, Together).
    ModelDetail:
      type: object
      description: Metadata and usage requirements for a model-provider pair.
      properties:
        id:
          type: string
        provider:
          type: string
    ChatCompletionRequest:
      type: object
      required:
        - model
        - messages
      properties:
        model:
          type: string
          description: Model identifier to run inference against.
        messages:
          type: array
          items:
            $ref: '#/components/schemas/ChatMessage'
        stream:
          type: boolean
          description: When true, stream the response as server-sent events.
    ChatMessage:
      type: object
      required:
        - role
        - content
      properties:
        role:
          type: string
          enum:
            - system
            - user
            - assistant
        content:
          type: string
    ChatCompletionResponse:
      type: object
      description: OpenAI-format chat completion.
      properties:
        id:
          type: string
        object:
          type: string
        created:
          type: integer
        model:
          type: string
        choices:
          type: array
          items:
            type: object
            properties:
              index:
                type: integer
              message:
                $ref: '#/components/schemas/ChatMessage'
              finish_reason:
                type: string
        usage:
          type: object
          properties:
            prompt_tokens:
              type: integer
            completion_tokens:
              type: integer
            total_tokens:
              type: integer
    Error:
      type: object
      properties:
        error:
          type: object
          properties:
            message:
              type: string
            type:
              type: string
            code:
              type: string