What GTFS data does caltrain-mcp use?
It uses official Caltrain GTFS data, including `stops.txt`, `trips.txt`, `stop_times.txt`, and `calendar.txt`, which are downloaded by running `fetch_gtfs.py`.
@davidyen1124
A Model Context Protocol (MCP) server for Caltrain schedules. Get real-time train departures using GTFS data, designed for Claude Desktop and other MCP clients.
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"caltrain": {
"command": "uvx",
"args": [
"caltrain-mcp"
]
}
}
}工具列表暂未提供
当前收录暂时没有可展示的工具列表,不代表服务没有工具。请查阅项目文档了解支持的能力。
点击上方按钮,在临时沙箱中运行此服务器并实时获取工具列表。
caltrain-mcp is a Model Context Protocol server that provides Caltrain schedule information using official GTFS data. It is intended for users of MCP-compatible clients (like Claude Desktop) who need to look up train departures, stations, and travel times between any two Caltrain stops.
Configure caltrain-mcp in your MCP client (e.g., Claude Desktop) by adding "caltrain": { "command": "uvx", "args": ["caltrain-mcp"] } to the MCP configuration file. The server can also be tested directly via Python imports or started with uvx caltrain-mcp. It relies on local GTFS data that must be fetched first by running fetch_gtfs.py.
when_iso parameterIt uses official Caltrain GTFS data, including `stops.txt`, `trips.txt`, `stop_times.txt`, and `calendar.txt`, which are downloaded by running `fetch_gtfs.py`.
Install the `uv` package manager, run `uv sync` to install dependencies, then download the GTFS data with `uv run python scripts/fetch_gtfs.py`. Finally, configure your MCP client to run `uvx caltrain-mcp`.
It exposes two tools: `next_trains(origin, destination, when_iso=None)` to get upcoming departures, and `list_stations()` to list all 31 Caltrain stations.
The server communicates via stdin/stdout using the Model Context Protocol (MCP). It does nothing when run directly in a terminal.
The README notes that actual arrival times may vary and that the server relies on the official GTFS data, so accuracy depends on that data. It handles mixed integer/string GTFS fields and supports 24+ hour time format.

Multiplayer coding in the agentic era through an MCP server for real-time codebase sync between your agents.

ListingGood AI Recommendation Engine — a hosted MCP server + Claude/Cursor skills that make Amazon's AI recommend your products. Free no-login scan: https://listinggood.com/scan Get API key: https://listinggood.com/deve
send-email is a remote MCP server available at https://next.hal9.ai.

Lease and contract records with ASC 842 lease accounting data: sites, leases, rent schedules, critical dates and journal entries, with confirmed write operations.

One photo of a building problem becomes a Triage Profile in about ten seconds: what the asset is, how urgent it is (EMERGENCY, TODAY, THIS WEEK or WHENEVER), which trade to call, and what to tell the resident. For proper
评论