# dispatchAI API Server

OpenAI-compatible inference API running on real phone hardware.

## Files

- api/gateway.py - FastAPI gateway (port 8081)
- api/phone_proxy_v2.py - Phone proxy (HTTP to ADB to llama-cli)
- api/mcp_server.py - MCP server (4 tools)
- api/test_api.py - Test script
- api/security_audit.py - Security audit
- api/Caddyfile - HTTPS config
- api/openapi.yaml - OpenAPI spec

## Quick Start

    pip install fastapi uvicorn httpx
    python -m uvicorn api.gateway:app --port 8081

## API

    POST /v1/chat/completions
    GET /v1/models
    GET /v1/usage
    GET /admin/phones