MCP Overview

The All Aboard Model Context Protocol (MCP) server gives any MCP-compatible AI assistant direct access to the API documentation and the GraphQL schema — so it can help you build integrations without you having to paste in docs or schema definitions manually.

What it provides

The MCP server exposes five tools to the model:

ToolDescription
list_documentationLists all available documentation pages with descriptions
read_documentationReads one or more documentation pages by slug
explore_schemaBrowses the GraphQL schema — queries, mutations, subscriptions, or types
get_typeGets full details of a specific GraphQL type, including fields and arguments
execute_queryExecutes a GraphQL query or mutation using your API key

See MCP Tools for the full reference.

Setup

The MCP server is hosted remotely and available at:

https://mcp.allaboard.eu/mcp
https://mcp.allaboard.eu/mcp

See the MCP Setup guide for instructions for your specific tool (Claude Code, Cursor, VS Code, OpenCode, and others).

Once the MCP server is connected, we recommend adding the following to your AGENTS.md, CLAUDE.md, or equivalent agent instruction file. This tells the model which tools are available and when to use them.

markdown
You have access to the All Aboard MCP server with tools for working with the
All Aboard European rail travel API.

## Tools

- **list_documentation** — Returns an index of all available documentation
  pages with slugs and short descriptions.
- **read_documentation** — Fetches one or more documentation pages by slug
  (comma-separated). Only fetch pages relevant to the current task.
- **explore_schema** — Lists queries, mutations, subscriptions, or types from
  the live GraphQL schema. Use this to find operation names and signatures.
- **get_type** — Returns full field and argument details for a specific
  GraphQL type. Always call this before writing a query that involves a type
  you haven't seen before.
- **execute_query** — Runs a GraphQL query or mutation against the API.
  Requires the user's API key. ALWAYS use the "test" environment unless the
  user explicitly asks for "live". Ask for the API key if not provided; it is 
  available from https://allaboard.eu/agent

Always call list_documentation at the start of an API-related task to orient
yourself. Then call read_documentation with only the slugs relevant to the
task. Call get_type to get details of a specific type before constructing the
query. Avoid fetching all pages at once.
You have access to the All Aboard MCP server with tools for working with the
All Aboard European rail travel API.

## Tools

- **list_documentation** — Returns an index of all available documentation
  pages with slugs and short descriptions.
- **read_documentation** — Fetches one or more documentation pages by slug
  (comma-separated). Only fetch pages relevant to the current task.
- **explore_schema** — Lists queries, mutations, subscriptions, or types from
  the live GraphQL schema. Use this to find operation names and signatures.
- **get_type** — Returns full field and argument details for a specific
  GraphQL type. Always call this before writing a query that involves a type
  you haven't seen before.
- **execute_query** — Runs a GraphQL query or mutation against the API.
  Requires the user's API key. ALWAYS use the "test" environment unless the
  user explicitly asks for "live". Ask for the API key if not provided; it is 
  available from https://allaboard.eu/agent

Always call list_documentation at the start of an API-related task to orient
yourself. Then call read_documentation with only the slugs relevant to the
task. Call get_type to get details of a specific type before constructing the
query. Avoid fetching all pages at once.