Request Interceptor

A man-in-the-middle HTTP proxy that logs requests/responses and provides special handling for AI API calls.

What it does

• Proxies HTTP requests to configurable targets
• Logs all requests and responses with timing data
• Detects AI API calls (OpenAI, Anthropic, OpenRouter, Azure) and extracts usage/cost data
• Provides a web dashboard for viewing logs
• Supports SSE streaming

Quick Start

docker compose up -d

• Dashboard: http://localhost:3100
• Proxy: http://localhost:3101

Default login: admin / changeme

Usage

Requests need a target. Three ways to specify it:

# 1. Query parameter
curl "http://localhost:3101/api/users?__target=https://api.example.com"

# 2. Header
curl -H "X-Target-URL: https://api.example.com" http://localhost:3101/api/users

# 3. Configure a default target or routing rules in the dashboard

AI API Example

curl "http://localhost:3101/v1/chat/completions" \
  -H "X-Target-URL: https://api.openai.com" \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-4o-mini", "messages": [{"role": "user", "content": "Hello"}]}'

The proxy will log the request, detect it as an AI call, and extract token usage and cost estimates.

Configuration

Environment variables:

Variable	Default	Description
PORT_ADMIN	3000	Dashboard port
PORT_PROXY	3001	Proxy port
ADMIN_USER	admin	Dashboard username
ADMIN_PASSWORD	changeme	Dashboard password
DATABASE_URL	file:/data/app.db	SQLite database path

Data Retention

• Request logs are deleted after 30 days
• Authorization headers are redacted after 3 days

Development

# Backend
npm install
npm run dev

# Frontend
cd frontend
npm install
npm run dev

License

MIT

---

_{Built by DPA+}