<2-3 sentences summarising what the project addresses and solves, the target audience, and the benefit.> Context+ is an MCP server designed for large-scale engineering, transforming massive codebases into searchable, hierarchical feature graphs. It targets developers demanding high accuracy by combining Retrieval Augmented Generation (RAG), Tree-sitter AST parsing, Spectral Clustering, and Obsidian-style linking, enabling semantic code understanding and navigation.

How It Works

The system leverages Tree-sitter for multi-language Abstract Syntax Tree (AST) parsing, enabling deep structural analysis. Spectral Clustering groups semantically related files, facilitating navigation through code by meaning rather than just text. Obsidian-style wikilinks connect features to code, creating a navigable "feature hub." Retrieval Augmented Generation (RAG) powered by Ollama embeddings allows for semantic code search, identifying code by intent. This integrated approach provides a powerful, hierarchical view of complex codebases.

Quick Start & Requirements

• Installation: No direct installation; runs via bunx or npx.
• IDE Integration: Configure via MCP settings in IDEs like Claude Code, Cursor, VS Code, Windsurf, and OpenCode.
• Prerequisites: Requires a running Ollama instance with specified embedding (nomic-embed-text) and chat (gemma2:27b or llama3.2) models. An OLLAMA_API_KEY is necessary.
• Commands: bunx contextplus to start the server, npx -y contextplus init [target] or bunx contextplus init [target] to generate IDE configuration files.
• Documentation: Configuration examples and setup commands are provided directly in the README.

Highlighted Details

• Comprehensive Tooling: Offers 11 MCP tools including semantic_code_search, get_blast_radius for impact analysis, run_static_analysis (TypeScript, Python, Rust, Go), and propose_commit with shadow restore points for safe AI-driven code changes.
• Architecture: Built with TypeScript over stdio using the Model Context Protocol (MCP) SDK, featuring a Core (AST parsing, embeddings, graph), Tools, and Git (shadow restore) layer.
• Runtime Cache: Utilizes a disk cache (.mcp_data/) for embeddings and a real-time tracker to incrementally refresh data on file changes, optimizing performance.
• Code Navigation: Features include structural AST trees, file skeletons showing API surfaces, and semantic navigation through clustered code.

Maintenance & Community

No specific details on contributors, sponsorships, or community channels were found in the provided text.

Licensing & Compatibility

No license information was specified in the provided text.

Limitations & Caveats

No explicit limitations, alpha status, or known bugs were mentioned in the provided text.