Give your AI agent
live transit vision
Metro-MCP connects LLMs and AI agents to current transit data from NYC Subway and DC Metro. Train positions, arrivals, delays—all streaming through MCP with instant updates.
1. Open Your Client's MCP/Connector settings
2. Add this URL: https://metro-mcp.anuragd.me/mcp
3. Click "Connect" and authorize via GitHub
4. Start asking questions about the Metro!
$ git clone https://github.com/Aarekaz/metro-mcp.git
$ cd metro-mcp && bun install
$ bunx wrangler kv namespace create "OAUTH_CLIENTS"
$ bunx wrangler secret put WMATA_API_KEY
$ bunx wrangler deploy
Powering transit-aware AI applications
From commuter assistants to analytics platforms, Metro-MCP provides the foundation for intelligent transit applications with up-to-date data and comprehensive coverage.
🚇
Live Arrival Predictions
Get current arrival times for any station. AI agents can answer "when's the next train?" with actual data, not guesses.
- Smart commute planning
- Travel assistants
- Delay notifications
📍
Train Position Tracking
Track every train in the system. See exactly where trains are, which direction they're heading, and train composition.
- System monitoring
- Crowding insights
- Route optimization
🚌
Multimodal Integration
Complete DC Metro bus support with instant positions, predictions, routes, and stops. Build comprehensive transit assistants.
- Bus + rail coordination
- Transfer planning
- Last-mile solutions
⚠️
Service Disruption Alerts
Stay ahead of disruptions. Get instant incident reports, elevator outages, and service advisories as they happen.
- Proactive route changes
- Accessibility updates
- Emergency notifications
🗺️
Rich Station Metadata
Complete station info with coordinates, transfer connections, and line information. Build spatial reasoning into your AI.
- Geographic queries
- Transfer optimization
- Accessibility info
🔄
Auto-Refresh Data Streams
Data updates every 7-10 seconds. Your AI always has the freshest information without manual polling.
- SSE event streams
- Efficient bandwidth
- Always up-to-date
Two major transit systems, one unified API
Comprehensive coverage of NYC Subway and DC Metro with different strengths and capabilities
| Feature | DC Metro (WMATA) | NYC Subway (MTA) |
|---|---|---|
| Train Arrivals | ✓ Live predictions | ✓ Live predictions |
| Train Positions | ✓ GPS tracking | ✓ GPS tracking |
| Service Alerts | ✓ Incidents & advisories | ✓ System-wide alerts |
| Bus Integration | ✓ Full support | – Rail only |
| Stations | 98 stations | 496 stations |
| Lines | 6 lines (Red, Orange, Silver, Blue, Yellow, Green) | 29 lines (A, C, E, 1, 2, 3, etc.) |
| Geographic Coverage | DC, MD, VA metro area | NYC 5 boroughs |
| Update Frequency | 7-10 seconds | 7-10 seconds |
Core MCP Tools
Essential tools that enable AI agents to access and query transit data effectively
Example queries
See how AI agents can answer transit questions naturally using Metro-MCP
The next Red Line train to Shady Grove arrives in 4 minutes at Metro Center. After that:
- 8 minutes — Second train
- 14 minutes — Third train
- Trains running on normal schedule, no reported delays
Current alerts affecting your route:
- Orange Line: Single tracking between Stadium-Armory and Minnesota Ave, expect 10-15 min delays
- Metro Center: Elevator to Red Line platform out of service, use escalator
- Consider: Blue Line alternative with 8-min transfer at L'Enfant Plaza
Fastest route (38 minutes total):
- Red Line to Metro Center (3 stops, 6 min)
- Transfer to Orange/Silver Line (2 min walk)
- Silver Line to Reagan National (6 stops, 28 min)
- Next departure in 5 minutes, arrive by 3:45 PM
Start building transit-aware AI
Deploy in minutes, scale infinitely. Metro-MCP runs on Cloudflare Workers with enterprise-grade reliability.
Built with modern technologies
TypeScript
Cloudflare Workers
MCP Protocol
WMATA API
MTA API