Hugging Face Info API

openapi: 3.2.0
info:
  title: Hugging Face Text Generation Inference Info API
  description: High-performance toolkit for deploying and serving large language models with optimized inference. Provides both a custom TGI API and an OpenAI-compatible Messages API for chat completions. Supports streaming, tool calling, structured output, grammar constraints, and multi-modal inputs.
  version: 1.0.0
  termsOfService: https://huggingface.co/terms-of-service
  contact:
    name: Hugging Face Support
    url: https://huggingface.co/support
  license:
    name: Apache 2.0
    url: https://www.apache.org/licenses/LICENSE-2.0
servers:
- url: https://api-inference.huggingface.co
  description: Hugging Face hosted TGI server
- url: http://localhost:8080
  description: Local TGI server (self-hosted)
security:
- bearerAuth: []
tags:
- name: Info
  description: Server and model information
paths:
  /v1/models:
    get:
      summary: List Available Models
      description: Get information about the model currently loaded on this TGI server.
      operationId: listModels
      tags:
      - Info
      security: []
      responses:
        '200':
          description: Available models
          content:
            application/json:
              schema:
                type: object
                properties:
                  object:
                    type: string
                    const: list
                  data:
                    type: array
                    items:
                      type: object
                      properties:
                        id:
                          type: string
                          description: Model ID
                        object:
                          type: string
                          const: model
                        created:
                          type: integer
                        owned_by:
                          type: string
              examples:
                Listmodels200Example:
                  summary: Default listModels 200 response
                  x-microcks-default: true
                  value:
                    object: example_value
                    data:
                    - id: abc123
                      object: example_value
                      created: 10
                      owned_by: example_value
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
  /info:
    get:
      summary: Get Server Information
      description: Get metadata about the TGI server including the loaded model, version, SHA, Docker label, and model-specific parameters.
      operationId: getInfo
      tags:
      - Info
      security: []
      responses:
        '200':
          description: Server and model information
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/Info'
              examples:
                Getinfo200Example:
                  summary: Default getInfo 200 response
                  x-microcks-default: true
                  value:
                    model_id: '500123'
                    model_sha: example_value
                    model_dtype: float16
                    model_device_type: example_value
                    model_pipeline_tag: example_value
                    max_concurrent_requests: 10
                    max_best_of: 10
                    max_stop_sequences: 10
                    max_input_length: 10
                    max_total_tokens: 10
                    waiting_served_ratio: 42.5
                    max_batch_total_tokens: 10
                    max_waiting_tokens: 10
                    validation_workers: 10
                    max_client_batch_size: 10
                    version: example_value
                    sha: example_value
                    docker_label: example_value
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
  /health:
    get:
      summary: Health Check
      description: Check the health of the TGI server. Returns 200 if the model is loaded and ready to serve requests.
      operationId: healthCheck
      tags:
      - Info
      security: []
      responses:
        '200':
          description: Server is healthy and ready
        '503':
          description: Server is not ready (model still loading)
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
  /metrics:
    get:
      summary: Get Prometheus Metrics
      description: Get server metrics in Prometheus format including request counts, latencies, queue sizes, and token throughput.
      operationId: getMetrics
      tags:
      - Info
      security: []
      responses:
        '200':
          description: Prometheus-formatted metrics
          content:
            text/plain:
              schema:
                type: string
              examples:
                Getmetrics200Example:
                  summary: Default getMetrics 200 response
                  x-microcks-default: true
                  value: example_value
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
  /tokenize:
    post:
      summary: Tokenize Text
      description: Tokenize a string of text using the model's tokenizer. Returns the token IDs and their string representations.
      operationId: tokenize
      tags:
      - Info
      requestBody:
        required: true
        content:
          application/json:
            schema:
              type: object
              required:
              - inputs
              properties:
                inputs:
                  type: string
                  description: Text to tokenize
                add_special_tokens:
                  type: boolean
                  default: true
                  description: Whether to add special tokens
            examples:
              TokenizeRequestExample:
                summary: Default tokenize request
                x-microcks-default: true
                value:
                  inputs: example_value
                  add_special_tokens: true
      responses:
        '200':
          description: Tokenized result
          content:
            application/json:
              schema:
                type: array
                items:
                  type: object
                  properties:
                    id:
                      type: integer
                      description: Token ID
                    text:
                      type: string
                      description: Token text
                    start:
                      type: integer
                    stop:
                      type: integer
                    special:
                      type: boolean
              examples:
                Tokenize200Example:
                  summary: Default tokenize 200 response
                  x-microcks-default: true
                  value:
                  - id: abc123
                    text: example_value
                    start: 10
                    stop: 10
                    special: true
      x-microcks-operation:
        delay: 0
        dispatcher: FALLBACK
components:
  schemas:
    Info:
      type: object
      properties:
        model_id:
          type: string
          description: Loaded model ID
          example: '500123'
        model_sha:
          type: string
          description: Model revision SHA
          example: example_value
        model_dtype:
          type: string
          description: Model data type
          enum:
          - float16
          - bfloat16
          - float32
          example: float16
        model_device_type:
          type: string
          description: Device type (cuda, cpu)
          example: example_value
        model_pipeline_tag:
          type: string
          description: Pipeline task tag
          example: example_value
        max_concurrent_requests:
          type: integer
          example: 10
        max_best_of:
          type: integer
          example: 10
        max_stop_sequences:
          type: integer
          example: 10
        max_input_length:
          type: integer
          description: Maximum input token length
          example: 10
        max_total_tokens:
          type: integer
          description: Maximum total tokens (input + output)
          example: 10
        waiting_served_ratio:
          type: number
          example: 42.5
        max_batch_total_tokens:
          type: integer
          example: 10
        max_waiting_tokens:
          type: integer
          example: 10
        validation_workers:
          type: integer
          example: 10
        max_client_batch_size:
          type: integer
          example: 10
        version:
          type: string
          description: TGI version
          example: example_value
        sha:
          type: string
          description: TGI build SHA
          example: example_value
        docker_label:
          type: string
          description: Docker image label
          example: example_value
  securitySchemes:
    bearerAuth:
      type: http
      scheme: bearer
      bearerFormat: HF Token
      description: Hugging Face API token or custom auth token for self-hosted instances.