@railway/mcp-server

Exposes Railway's core infrastructure operations through the Model Context Protocol, allowing LLM agents and Claude instances to programmatically query and manage Railway projects, services, deployments, and environments. Implements MCP server specification with Railway API client bindings, enabling structured tool calling for infrastructure automation without direct API knowledge.

Automatically generates MCP-compliant tool schemas (JSON Schema format) from Railway API endpoints, mapping REST operations to structured function definitions that Claude and other LLM clients can invoke. Implements schema generation patterns that translate Railway API parameters, response types, and error codes into MCP tool specifications with proper type hints and validation.

Manages Railway API authentication tokens within the MCP server context, accepting API credentials at server initialization and securely passing them to all Railway API calls. Implements credential handling patterns that keep tokens out of tool parameters (preventing exposure in LLM logs) while ensuring they're available to all downstream API operations.

Provides tools to query current deployment status (running, failed, building, etc.) and detect changes since last query, enabling LLM agents to monitor Railway deployments without continuous polling. Implements state tracking patterns that cache deployment metadata and compare against fresh API queries to identify status transitions, new errors, or completed builds.

Exposes Railway's environment variable and secret management APIs through MCP tools, allowing Claude to query, create, update, and delete environment variables across Railway services and environments. Implements secure parameter passing patterns that prevent secrets from being logged or exposed in tool parameters, using server-side secret handling instead.

Provides tools to discover and introspect Railway services, plugins, and their configurations within a project, returning metadata about available services, their ports, environment variables, and dependencies. Implements introspection patterns that query Railway's project structure and return structured metadata that Claude can use to understand the deployment topology.

Exposes Railway's deployment and service logs through MCP tools, allowing Claude to retrieve historical logs or stream real-time logs for debugging and monitoring. Implements log retrieval patterns that fetch logs from Railway's log storage and format them for LLM consumption, with optional filtering by service, environment, or time range.

Provides MCP tools to trigger new deployments, redeploy specific versions, and rollback to previous deployments. Implements deployment orchestration patterns that queue deployment requests with Railway's build system and track deployment progress, enabling Claude to automate deployment workflows and recovery procedures.