Bart

@bart what's the next train from Powell Street to SFO Airport leaving after 2:30 PM today?

Based on the current schedule, the next BART train from Powell Street (POWL) to SFO Airport (SFIA) departing after 2:30 PM is at 2:38 PM on the Yellow Line. The journey takes approximately 32 minutes with one transfer at Balboa Park, arriving at SFO at 3:10 PM. The following departure is at 2:53 PM, arriving at 3:25 PM. Both routes require a transfer to the Yellow Line for airport service.

This example uses the schedule depart tool to plan a trip with time constraints. Note that responses reflect published schedules, not real-time delays or service changes. Users must provide valid station names or codes—the MCP can look up codes if needed. Schedule data updates periodically but may not reflect same-day service advisories.

@bart download the latest GTFS feed and tell me which routes serve the Embarcadero station on weekdays

I've downloaded the current GTFS Static Schedule feed (google_transit.zip). Analyzing the stops and routes data: Embarcadero station (EMBR) is served by five routes on weekdays: Red Line (Richmond–Daly City/Millbrae), Yellow Line (Antioch–SFO/Millbrae), Green Line (Berryessa–Daly City), Blue Line (Dublin/Pleasanton–Daly City), and Orange Line (Richmond–Berryessa). Peak service runs every 4-6 minutes during commute hours, with 12-15 minute headways midday.

This example demonstrates downloading raw GTFS data for deeper analysis beyond the API's structured endpoints. The GTFS feed is a 5-15 MB zip file containing CSV tables—useful for bulk analysis or custom integrations. The MCP can parse this data to answer complex routing questions, but processing large feeds may take 10-20 seconds.

@bart check real-time trip updates and summarize any current delays affecting trains to Oakland

Checking GTFS-RT trip updates feed: Currently, two delays are affecting Oakland-bound service. The 3:15 PM Yellow Line train from SFO is running 8 minutes late due to equipment issues at San Bruno. The 3:22 PM Orange Line from Richmond has a 4-minute delay at 19th Street Oakland. All other Oakland-bound trains (Red, Green, Blue lines) are operating on schedule. No service advisories are active for transbay tube or East Bay stations.

This example uses the real-time GTFS-RT feed to surface live service disruptions. The MCP returns raw protobuf data that the AI interprets into readable summaries. Real-time data refreshes every 30-60 seconds but may lag during major incidents. This requires understanding BART's route structure—the AI can cross-reference with static schedule data to provide context.

When BART MCP helps remote teams coordinate SF office days

A 6-person engineering team in Oakland schedules quarterly in-person sprints at their SF office. The ops lead uses Switchy with the BART MCP to pull live departure schedules and GTFS trip updates the morning of each sprint kickoff, then shares arrival windows in Slack so no one waits 20 minutes at Embarcadero. The MCP's 10 tools cover schedule lookups and real-time feeds, but this only works if your team actually rides BART—if half the team drives or takes Caltrain, you're maintaining two separate transit contexts. The buying call: if 3+ people on your team use BART for the same destination more than once a month, this MCP turns commute coordination from a 12-message thread into a single AI query.

BART MCP wins for small support teams answering schedule questions

A 4-person support team at a Bay Area mobility startup fields 40–60 tickets a week asking 'why is my trip delayed' or 'when does the next train leave Fremont.' They connect the BART MCP to Switchy and let the AI pull GTFS-RT trip updates and route info in real time, cutting average response time from 8 minutes to under 2. The MCP's API key auth is simple enough that the support lead set it up in one afternoon. The threshold: if your ticket volume drops below 20/week or your app covers more than just BART, the MCP becomes narrow overhead. This works when BART questions are 60%+ of your inbound volume and your team is too small to justify a custom integration.

When BART MCP helps community organizers time venue access

A 3-person team runs monthly developer meetups in downtown SF and needs to tell 80–120 attendees when doors open relative to train arrivals. They use Switchy with the BART MCP to query departure schedules from the 5 most common origin stations, then auto-generate a 'how to get here' doc that accounts for evening service gaps. The MCP's schedule-depart and schedule-arrive tools let them model trips in both directions without opening the BART website. The trade-off: this only pays off if your event is transit-accessible and you're running it more than twice a quarter—one-off events don't justify the setup. If your attendees mostly drive or your venue is outside the BART footprint, skip this and use Google Maps links instead.