Sign inTry it

FlixBus MCP server: let your AI agent search bus trips between two cities

Wield is an MCP server that gives your AI agent one FlixBus tool, flixbus.de/search-trips. The tool searches bus trips between two cities on a date. It returns departure and arrival times, stations, duration, transfers, seats left and the shop price, including the platform fee. A link opens the same search in the FlixBus shop, where you book.

Updated

Live recipeflixbus.de/search-tripsv1

FlixBus Search bus trips between two cities on a date

Read-only, no loginPasses the daily check

Parameters (5)
adultsinteger
datestringrequiredDeparture date YYYY-MM-DD
fromstringrequiredDeparture city, e.g. 'Berlin'
only_availablebooleanDrop sold-out and otherwise unbookable trips
tostringrequiredArrival city, e.g. 'München'
Use it with your agent
Measured on the live site52× faster
MetricWield recipeBrowser Use + Claude Sonnet 5.5
Success100%100%
Median time791 ms41 s
Cost per run$0$0.227

9 runs each, 7 November 2026. How we measure

What your agent gets

RecipeWhat it doesLoginStatus (7 Nov 2026)
flixbus.de/search-tripsSearches bus and train trips between two cities on a date, with times, stations, transfers, seats and priceNoOK, 7 of 7 checks, about 1 s

The tool takes from and to (city names), date (YYYY-MM-DD) and optional adults and only_available. City names are free text. The tool uses the first match and returns the city it used, so your agent can check it. By default, sold-out and unbookable trips are dropped.

A typical run makes three requests. In our check on 7 November 2026, the search finished in about one second.

How a request flows

  1. You ask your agent: "Find FlixBus trips from Berlin to München on 20 November 2026 for two adults."
  2. The agent calls find_recipe, then run with flixbus.de/search-trips and the date.
  3. wield-mcp passes the run through the local bridge to the Wield extension.
  4. The extension sends the requests from a flixbus.de tab, so the search runs as your browser would run it.
  5. The trips go back to your agent. Nothing passes through wield.run.

Your agent can then open the shop_url from the result to book in the FlixBus shop.

What leaves your computer

The only host is global.api.flixbus.com. A search makes three requests, and the requests are sent without your FlixBus cookies. Booking stays in the FlixBus shop, so Wield never books, pays or enters your details.

Wield logs each request on your computer only: method, host, path, status and timing. It never logs bodies or query values. Telemetry is off by default. Wield never reads, copies or sends your cookies or passwords.

FlixBus terms, in plain words

We have not reviewed FlixBus's terms as a lawyer would, so this is not legal advice. Our reading is that FlixBus's terms restrict automated extraction of timetable and price data. This recipe runs one search that you start, from your own browser, with three requests. It does not collect timetables in bulk and does not book. Check FlixBus's terms yourself before you automate anything else.

Limits and what is not working yet

  • Read only. Booking happens in the FlixBus shop through shop_url.
  • City names are matched, not confirmed. The first match wins, so check the from and to values in the result.
  • Sold-out trips are hidden. only_available is true by default.
  • Prices include the platform fee. The result also carries the price without it.
  • Run limits apply. A run has at most 12 steps, 40 requests, 5 MB per response and 60 seconds.

FAQ

Is there an official FlixBus MCP server?

Not that we know of as of 7 November 2026.

Which AI apps work with it?

Any MCP client. npx -y wield-mcp init sets up Claude Desktop, Claude Code, Cursor and Codex. Other MCP clients work with the standard stdio config.

Does it need a FlixBus login?

No. Search works without an account.

How fast is it compared with a browser agent?

The benchmark task checks for at least one trip on the requested date with departure, arrival and price. As of 7 November 2026, and still preliminary, the Wield recipe passed 9 of 9 runs with a median of 791 ms and no model cost. Browser Use with Claude Sonnet 5.5 also passed its runs, 3 of 3, with a median of 41.1 seconds and $0.23 per run. The gap here is speed and cost, not success. See the FlixBus benchmark.