FlixBus
Plan a trip with FlixBus directly in ChatGPT. Search routes between supported destinations across the globe, browse available bus and train options for your chosen date and proceed directly to checkout.
- Homepage
- https://global.flixbus.com
- Remote URL
https://mcp-gpt.dist.flix.tech/mcp- Auth
- NONE
Tools (2)
Extracted live via the executor SDK.
-
prepare_trip_searchResolve free-text origin and destination cities for a FlixBus, FlixTrain, Greyhound or Kâmil Koç trip search and return canonical city IDs, country codes, transport preference, and route-valid passenger guidance. Use it when the user gives city names rather than canonical IDs. It can usually handle minor spelling differences, diacritics, and common variants without restating the city, but if you notice a misspelling yourself, send this tool a corrected version. Treat route passenger-rule warnings as informational for generic requests, but set passengerSignals when the user explicitly mentions children, youth travelers, or lap infants so the tool can require follow-up where the route needs more detail. Ask a follow-up question only when the result explicitly says disambiguation or passenger clarification is required. If the result is ready for search, call search_trips directly without sending an extra assistant confirmation message in between; rely on the tool invocation status text instead. After success, call search_trips once (not several parallel calls for different times of day). Never expose internal identifiers such as type keys or reason enums in user-facing text.
-
search_tripsSearch live FlixBus, FlixTrain, Greyhound or Kâmil Koç options for a specific origin, destination, and departure date, and render the widget from the fresh result. Do not call this tool unless you have preferredLocale, canonical city IDs, exact departureDate, transportPreference, and a complete passenger selection. Use route-valid passenger types. Set transportPreference to
trainfor train-only requests,busfor bus-only requests, andanyotherwise; the server enforces that filter. Set directOnly totruewhen the user asks for direct trips only or says no transfers/connections. Use sortBy for deterministic ranking (departure_asc,price_asc,duration_asc,transfer_count_asc). Use departureTimeRange to filter by local departure times with exactHH:MMbounds, for example{ start: "11:00", end: "23:30" }. The range is inclusive and may cross midnight whenstartis later thanend. When the user does not specify a time slot, the tool defaults to daytime departures from07:00to22:00. If that window is empty or nearly empty, the server may broaden results automatically instead of returning an empty widget. Never issue multiple parallel calls that only differ by time filters to "cover" the day. Resolve relative dates liketoday,tomorrow, orthis Fridayto an exact YYYY-MM-DD date before calling. Keep at least one actual passenger in the selection because bike_slot and wheelchair cannot be booked on their own, and keep actual passengers at 10 or fewer. Use the same preferredLocale throughout the booking flow, choosing the best supported commercial locale from the user's language and clear regional context. When chaining from prepare_trip_search, do not send a separate assistant progress message before this call; rely on the tool invocation status text. Use this for each user-requested change so the widget reflects current prices and availability, not stale prior results. Do not use it for wide date sweeps or speculative price hunting.