Fastly · Capability

Fastly AI Accelerator — Chat Completions

Fastly AI Accelerator semantic-caching proxy for LLM chat completions. Provides OpenAI- and Google Gemini-compatible endpoints served from the Fastly edge, returning cached responses for semantically similar prompts.

Fastly AI Accelerator — Chat Completions is a Naftiko capability published by Fastly, one of 73 capabilities the APIs.io network indexes for this provider. It bundles 3 operations across the POST method.

The capability includes 3 state-changing operations. Lead operation: Create OpenAI-compatible chat completion via Fastly AI Accelerator semantic cache. Can be deployed as a REST endpoint, MCP tool, or Agent Skill via Naftiko.

Tagged areas include Fastly, AI Accelerator, AI, LLM, and Semantic Caching.

Run with Naftiko FastlyAI AcceleratorAILLMSemantic Caching

What You Can Do

POST
Createopenaichatcompletion — Create chat completion via OpenAI
/v1/openai/v1/chat/completions
POST
Generategeminicontent — Generate content via Google Gemini
/v1/gemini/v1/models/{model}/generate-content
POST
Createembeddings — Create embeddings
/v1/openai/v1/embeddings

MCP Tools

openai-chat-completion

Create OpenAI-compatible chat completion via Fastly AI Accelerator semantic cache

gemini-generate-content

Generate Google Gemini content via Fastly AI Accelerator

openai-embeddings

Create OpenAI embeddings via Fastly AI Accelerator

idempotent

Capability Spec

ai-accelerator-chat.yaml Raw ↑ 
naftiko: 1.0.0-alpha2
info:
  label: Fastly AI Accelerator — Chat Completions
  description: Fastly AI Accelerator semantic-caching proxy for LLM chat completions. Provides OpenAI- and Google Gemini-compatible endpoints served from the Fastly edge, returning cached responses for semantically similar prompts.
  tags:
  - Fastly
  - AI Accelerator
  - AI
  - LLM
  - Semantic Caching
  created: '2026-05-22'
  modified: '2026-05-22'
binds:
- namespace: env
  keys:
    FASTLY_API_KEY: FASTLY_API_KEY
capability:
  consumes:
  - type: http
    namespace: ai-accelerator
    baseUri: https://api.fastly.ai
    description: Fastly AI Accelerator semantic-caching proxy for LLM chat completions.
    resources:
    - name: openai-v1-chat-completions
      path: /openai/v1/chat/completions
      operations:
      - name: createopenaichatcompletion
        method: POST
        description: Create chat completion via OpenAI (semantically cached)
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: gemini-v1-models-model-generate-content
      path: /gemini/v1/models/{model}:generateContent
      operations:
      - name: generategeminicontent
        method: POST
        description: Generate content via Google Gemini (semantically cached)
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: openai-v1-embeddings
      path: /openai/v1/embeddings
      operations:
      - name: createembeddings
        method: POST
        description: Create embeddings via OpenAI
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    authentication:
      type: apikey
      key: Fastly-Key
      value: '{{env.FASTLY_API_KEY}}'
      placement: header
  exposes:
  - type: rest
    namespace: ai-accelerator-rest
    port: 8080
    description: REST adapter for Fastly AI Accelerator.
    resources:
    - path: /v1/openai/v1/chat/completions
      name: openai-v1-chat-completions
      description: REST surface for OpenAI-compatible chat completions through AI Accelerator.
      operations:
      - method: POST
        name: createopenaichatcompletion
        description: Create chat completion via OpenAI
        call: ai-accelerator.createopenaichatcompletion
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/gemini/v1/models/{model}/generate-content
      name: gemini-v1-models-model-generate-content
      description: REST surface for Google Gemini generateContent through AI Accelerator.
      operations:
      - method: POST
        name: generategeminicontent
        description: Generate content via Google Gemini
        call: ai-accelerator.generategeminicontent
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/openai/v1/embeddings
      name: openai-v1-embeddings
      description: REST surface for OpenAI embeddings through AI Accelerator.
      operations:
      - method: POST
        name: createembeddings
        description: Create embeddings
        call: ai-accelerator.createembeddings
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: ai-accelerator-mcp
    port: 9090
    transport: http
    description: MCP adapter for Fastly AI Accelerator.
    tools:
    - name: openai-chat-completion
      description: Create OpenAI-compatible chat completion via Fastly AI Accelerator semantic cache
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: ai-accelerator.createopenaichatcompletion
      outputParameters:
      - type: object
        mapping: $.
    - name: gemini-generate-content
      description: Generate Google Gemini content via Fastly AI Accelerator
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: ai-accelerator.generategeminicontent
      outputParameters:
      - type: object
        mapping: $.
    - name: openai-embeddings
      description: Create OpenAI embeddings via Fastly AI Accelerator
      hints:
        readOnly: false
        destructive: false
        idempotent: true
      call: ai-accelerator.createembeddings
      outputParameters:
      - type: object
        mapping: $.