Drupal Scout MCP
AI-powered Drupal development assistant.
A Model Context Protocol (MCP) server that brings Drupal expertise to AI assistants like Claude. It helps developers by providing instant access to Drupal documentation, best practices, and code examples directly in their AI conversations. Makes Drupal development faster and more efficient.
The problem
Drupal developers using AI tools constantly context-switch between the codebase and the live database to answer simple questions. What content types exist? Which modules depend on this one? What is in watchdog? Each answer costs a drush call, a grep, a YAML inspection, and a few hundred tokens. Drupal Scout MCP collapses all of that into one MCP server.
The approach
A Python MCP server (FastMCP) with eleven tool modules. Static indexing parses .info.yml, .services.yml, and .routing.yml for offline insights. Live drush php:eval queries pull entities, fields, and views from active config. The result is a hybrid that answers from code when speed matters and from the database when accuracy matters.
Notable decisions
- Lazy initialization. Pre-indexing deferred to first request so MCP startup never times out.
- Read-only by design. The tools query; the AI executes drush or composer for changes.
- CSV export for large entity dumps. Avoids token-budget blowups on big sites.
- Drupal.org API integration so AI can search the contrib ecosystem without leaving the chat.
Status
Published on PyPI as drupal-scout-mcp. Active development with recent work on Excel export, taxonomy improvements, and CI.
Want similar work for your team?
AI-native senior engineering for AI, cloud, and web. Open to consulting engagements and full-time roles.