Aleph Alpha Models API

openapi: 3.2.0
info:
  title: Aleph Alpha Models API
  version: '1.0'
  description: 'Operations tagged models across 3 of this provider''s published API definitions: aleph-alpha-pharia-inference-openapi.json, aleph-alpha-pharia-studio-openapi.json, aleph-alpha-responses-openapi.json. Each path carries the servers of the definition it was published in.'
servers:
- url: '{host}/v1'
  variables:
    host:
      default: https://api.pharia.example.com
- url: '{host}/v1/studio'
  variables:
    host:
      default: https://api.pharia.example.com
- url: '{host}'
  variables:
    host:
      default: https://api.pharia.example.com
tags:
- name: models
paths:
  /models/{modelName}/tokenizer:
    parameters:
    - name: modelName
      in: path
      description: Name of the model
      required: 'true'
      schema:
        type: string
    servers:
    - url: '{host}/v1'
      variables:
        host:
          default: https://api.pharia.example.com
    get:
      operationId: getModelTokenizer
      summary: Get the tokenizer of a model
      description: Returns a representation of the tokenizer that was used to train that model
      tags:
      - models
      security:
      - token: []
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                type: object
  /models_available:
    servers:
    - url: '{host}/v1'
      variables:
        host:
          default: https://api.pharia.example.com
    get:
      summary: Available models (Deprecated)
      deprecated: 'true'
      description: Will return all currently available models.
      operationId: availableModels
      tags:
      - models
      security:
      - token: []
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                type: array
                items:
                  properties:
                    name:
                      type: string
                    description:
                      type: string
                    hostings:
                      type: array
                      items:
                        type: string
  /model-settings:
    servers:
    - url: '{host}/v1'
      variables:
        host:
          default: https://api.pharia.example.com
    get:
      summary: Settings of models
      description: Only models available to the client will be listed.
      operationId: modelSettings
      tags:
      - models
      security:
      - token: []
      responses:
        '200':
          description: OK
          content:
            application/json:
              schema:
                type: array
                items:
                  properties:
                    name:
                      type: string
                    status:
                      type: string
                      enum:
                      - available
                      - unavailable
                      description: "The current availability status of the model. Currently supported states are:\n  - `available`: The model is configured and a matching worker is connected to serve it.\n  - `unavailable`: The model is configured but no worker has shown recent activity to serve it.\n"
                    description:
                      type: string
                    multimodal:
                      type: boolean
                      description: Feature flag for whether or not multimodal prompts are available to users.
                    max_context_size:
                      type: number
                      description: The maximum context size of this model.
                    semantic_embedding:
                      type: boolean
                      description: True if this model supports semantic embeddings.
                    completion_type:
                      type: string
                      enum:
                      - none
                      - full
                      description: 'The completion type supported by the model.


                        - `none`: The model has not been trained to support completions. Trying to trigger a completion request will lead to a validation error.

                        - `full`: The model has been trained to support completions.

                        '
                    embedding_type:
                      type: string
                      enum:
                      - none
                      - raw
                      - semantic
                      - instructable
                      description: 'The embedding type supported by the model.


                        - `none`: The model cannot be used for embeddings. The scheduler will reject requests for embeddings to this model.

                        - `raw`: The model has not explicitly been trained to support embeddings. However, it is possible to retrieve the embedding details technically. This option maps to the `/embed` endpoint.

                        - `semantic`: The model has been trained with a switchable set of weights usable for semantic embedding retrieval. This option maps to the `/semantic_embed` endpoint.

                        - `instructable`: The model has been trained to support any custom instruction for embedding retrieval. This option maps to the `/instructable_embed` endpoint.


                        This flag replaces `semantic_embedding_enabled` and should always be set. If `embedding_type` is unset, `semantic_embedding_enabled` will take control. If both `embedding_type` and `semantic_embedding_enabled` are used, implausible combinations are rejected.

                        '
                    aligned:
                      type: boolean
                      description: Specifies whether the model is aligned s.t. end users can be warned about the model's limitations.
                    chat:
                      type: boolean
                      description: True if this model is supported by the chat endpoint.
                    prompt_template:
                      type: string
                      description: A prompt template that can be used for this model.
                    maximum_completion_tokens:
                      nullable: 'true'
                      type: number
                      description: Maximum number of tokens that can be generated by a user request. Not to be confused with the context size which limits the request size including the prompt.
                    model_card:
                      nullable: 'true'
                      type: string
                      format: uri
                      description: Link to further information about the model
  /models:
    servers:
    - url: '{host}/v1/studio'
      variables:
        host:
          default: https://api.pharia.example.com
    get:
      tags:
      - models
      summary: Get All Model Cards
      description: Get all `ModelCard` from the list of models that are available.
      operationId: get_all_model_cards_models_get
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema:
                items:
                  $ref: '#/components/schemas/ModelCardResponse'
                type: array
                title: Response Get All Model Cards Models Get
      security:
      - BearerAuth: []
  /models/complete:
    servers:
    - url: '{host}/v1/studio'
      variables:
        host:
          default: https://api.pharia.example.com
    post:
      tags:
      - models
      summary: Complete
      description: Send a streamed complete request against the inference api. This always sets `stream=True` in the request body.
      operationId: complete_models_complete_post
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ModelCompleteRequest'
        required: true
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema: {}
        '422':
          description: Validation Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/HTTPValidationError'
      security:
      - BearerAuth: []
  /models/chat/completions:
    servers:
    - url: '{host}/v1/studio'
      variables:
        host:
          default: https://api.pharia.example.com
    post:
      tags:
      - models
      summary: Chat Completion
      operationId: chat_completion_models_chat_completions_post
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ChatCompletionRequest'
        required: true
      responses:
        '200':
          description: Successful Response
          content:
            application/json:
              schema: {}
        '422':
          description: Validation Error
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/HTTPValidationError'
      security:
      - BearerAuth: []
components:
  schemas:
    Message:
      properties:
        role:
          type: string
          title: Role
        content:
          type: string
          title: Content
      type: object
      required:
      - role
      - content
      title: Message
      description: Chat message model.
    ModelCardResponse:
      properties:
        name:
          type: string
          title: Name
          description: The name of the model
          examples:
          - luminous-base-control
          - llama-3.1-70b-instruct
        status:
          type: string
          enum:
          - available
          - unavailable
          title: Status
          description: The current availability status of the model. Currently supported states are 'available' and 'unavailable'
          examples:
          - available
          - unavailable
        description:
          anyOf:
          - type: string
          - type: 'null'
          title: Description
          description: A brief description of the model's capabilities and features
          examples:
          - A model that can generate text
        max_context_size:
          type: integer
          title: Max Context Size
          description: The maximum number of tokens the model can process in a single input
          examples:
          - 2048
          - 8192
        aligned:
          type: boolean
          title: Aligned
          description: ''
        semantic_embedding:
          type: boolean
          title: Semantic Embedding
          description: Whether the model can generate semantic embeddings
        worker_type:
          type: string
          title: Worker Type
          description: Type of worker that serves the model. Can either be 'luminous' or 'vllm'. If 'luminous' is set, the model supports advanced completion parameters. If called with these parameters and it is not supported, it will raise an error. This is being deprecated because the Inference API does not expose what worker the model is running on anymore.
          deprecated: true
          examples:
          - vllm
          - luminous
          - ''
        multimodal:
          type: boolean
          title: Multimodal
          description: Whether the model can process multiple types of input data (e.g., text, images)
        chat:
          type: boolean
          title: Chat
          description: Whether the model is supported by the chat endpoint
        completion_type:
          type: string
          enum:
          - full
          - none
          title: Completion Type
          description: The completion type supported by the model. It states if the model has not been trained to support completions.
          examples:
          - full
          - none
        prompt_template:
          type: string
          title: Prompt Template
          description: The prompt template that should be used to prompt the model
          examples:
          - "{{instruction}} \n #Output: "
        category:
          anyOf:
          - type: string
          - type: 'null'
          title: Category
          description: The category of the model
          examples:
          - TextGeneration
        link:
          type: string
          title: Link
          description: The link to the model card
          examples:
          - https://api.aleph-alpha.com
        maximum_completion_tokens:
          anyOf:
          - type: integer
          - type: 'null'
          title: Maximum Completion Tokens
          description: The maximum number of tokens the model will provide for the output.
          examples:
          - 2048
          - 8192
      type: object
      required:
      - name
      - status
      - description
      - max_context_size
      - aligned
      - semantic_embedding
      - worker_type
      - multimodal
      - chat
      - completion_type
      - prompt_template
      - link
      title: ModelCardResponse
      description: Response model for model cards.
    ChatCompletionRequest:
      properties:
        messages:
          items:
            $ref: '#/components/schemas/Message'
          type: array
          title: Messages
        model:
          type: string
          title: Model
        stream:
          type: boolean
          title: Stream
          default: true
      additionalProperties: true
      type: object
      required:
      - messages
      - model
      title: ChatCompletionRequest
      description: Request model for chat completion.
    ValidationError:
      properties:
        loc:
          items:
            anyOf:
            - type: string
            - type: integer
          type: array
          title: Location
        msg:
          type: string
          title: Message
        type:
          type: string
          title: Error Type
        input:
          title: Input
        ctx:
          type: object
          title: Context
      type: object
      required:
      - loc
      - msg
      - type
      title: ValidationError
    ModelCompleteRequest:
      properties:
        model:
          type: string
          title: Model
        prompt:
          type: string
          title: Prompt
      additionalProperties: true
      type: object
      required:
      - model
      - prompt
      title: ModelCompleteRequest
      description: Request model for model completion.
    HTTPValidationError:
      properties:
        detail:
          items:
            $ref: '#/components/schemas/ValidationError'
          type: array
          title: Detail
      type: object
      title: HTTPValidationError
  securitySchemes:
    token:
      type: http
      scheme: bearer
      description: Can be generated in your [Aleph Alpha profile](https://app.aleph-alpha.com/profile)
    BearerAuth:
      type: http
      scheme: bearer
      bearerFormat: JWT
    HTTPBearer:
      type: http
      scheme: bearer
x-refined-from:
- aleph-alpha-pharia-inference-openapi.json
- aleph-alpha-pharia-studio-openapi.json
- aleph-alpha-responses-openapi.json
x-tagGroups:
- name: PhariaData API
  tags:
  - Repositories
  - Datasets
  - Stages
  - Downloads
  - Transformations
  - Connectors
- name: Pharia Search API
  tags:
  - Search Store
  - Document