Skip to main content

Glossary

Key terms used throughout the MCP ADR Analysis Server documentation.

ADR (Architectural Decision Record)​

A short document that captures an important architectural decision, including the context, the decision itself, and its consequences. ADRs provide a log of decisions that affect a project's structure and behavior. This project follows the Nygard template.

AST (Abstract Syntax Tree)​

A tree representation of the syntactic structure of source code. The server uses AST analysis (via tree-sitter) to understand code patterns, detect architectural decisions embedded in code, and perform intelligent code linking.

Firecrawl​

An optional web scraping and content extraction service that enhances the server's research capabilities. When enabled, it allows tools like llm_web_search to gather real-time information from the web for more comprehensive ADR generation. See the Firecrawl Setup Guide.

Knowledge Graph​

An in-memory data structure maintained by the server that models relationships between ADRs, code implementations, technology decisions, and their impacts. It powers intelligent features like pattern recognition and cross-ADR dependency analysis.

MCP (Model Context Protocol)​

An open protocol created by Anthropic that enables AI assistants to connect to external tools and data sources. MCP defines a standard way for clients (like Claude Desktop) to discover and invoke tools provided by servers (like this one). See the official specification.

OpenRouter​

An API gateway that provides access to multiple AI models (Claude, GPT, Gemini, etc.) through a single API key. The server uses OpenRouter for AI-powered analysis in "full" execution mode. Sign up at openrouter.ai.

Ripgrep​

A fast text search tool (rg) used by the server for code search operations such as Smart Code Linking. It searches file contents significantly faster than traditional grep. The server uses it alongside fast-glob for file matching.

tree-sitter​

An incremental parsing library that builds and maintains ASTs for source code. The server uses tree-sitter to analyze code in multiple languages (TypeScript, Python, Java, Go, etc.) without executing it, enabling language-aware architectural analysis.