# Generated by API Evangelist (build-phrasing.py). Our phrasing, not observed demand.
overlay: 1.0.0
info:
  title: API Evangelist conversational phrasing for Hugging Face Text Generation Inference Info API
  version: 1.0.0
extends: openapi/hugging-face-info-api-openapi.yml
actions:
- target: $.info
  update:
    x-apievangelist-phrasing:
      method: generated
      generated: '2026-10-01'
      generator: build-phrasing.py
      label: Generated by API Evangelist
      operations: 5
- target: $.paths['/v1/models'].get
  update:
    x-apievangelist-phrasing:
      intent: Show the model loaded on a TGI server
      effect: read
      questions:
      - Which model is currently loaded on my Text Generation Inference server?
      - Can I query a TGI deployment's OpenAI-style models list?
      instructions:
      - text: List the models available on this TGI server.
      - text: Tell me which model the TGI deployment has loaded.
      method: generated
      generated: '2026-10-01'
- target: $.paths['/info'].get
  update:
    x-apievangelist-phrasing:
      intent: Get TGI server version and model metadata
      effect: read
      questions:
      - What version, commit SHA and Docker label is my TGI server running?
      - Can I read the max input length and other model parameters the server was started with?
      instructions:
      - text: Get the server information from TGI.
      - text: Show me the TGI version and model configuration details.
      method: generated
      generated: '2026-10-01'
- target: $.paths['/health'].get
  update:
    x-apievangelist-phrasing:
      intent: Check whether a TGI server is ready
      effect: read
      questions:
      - Is my TGI server up with the model loaded and ready to serve requests?
      - Which probe should a load balancer hit to confirm the text generation server is healthy?
      instructions:
      - text: Check the health of the TGI server.
      - text: Confirm the text generation server is ready for traffic.
      method: generated
      generated: '2026-10-01'
- target: $.paths['/metrics'].get
  update:
    x-apievangelist-phrasing:
      intent: Scrape TGI Prometheus metrics
      effect: read
      questions:
      - Can I scrape queue size and token throughput from a TGI server with Prometheus?
      - Where are the request count and latency metrics for Text Generation Inference exposed?
      instructions:
      - text: Fetch the Prometheus metrics from the TGI server.
      - text: Show me TGI queue size and token throughput metrics.
      method: generated
      generated: '2026-10-01'
- target: $.paths['/tokenize'].post
  update:
    x-apievangelist-phrasing:
      intent: Tokenize text with the loaded model's tokenizer
      effect: read
      questions:
      - How many tokens does a piece of text become with the model loaded on TGI?
      - Can I see token IDs and their string pieces, with or without special tokens?
      instructions:
      - text: Tokenize {inputs} with the loaded model.
        slots:
          inputs: requestBody.inputs
      - text: Split {inputs} into tokens, setting add special tokens to {add_special_tokens}.
        slots:
          inputs: requestBody.inputs
          add_special_tokens: requestBody.add_special_tokens
      method: generated
      generated: '2026-10-01'