contextplus

Overview

Context+ is a sophisticated MCP server designed for developers working with large-scale codebases. It combines multiple advanced techniques including RAG (Retrieval-Augmented Generation), Tree-sitter AST parsing, Spectral Clustering, and Obsidian-style linking to create a searchable, hierarchical feature graph. The server provides 17 tools across categories including discovery, analysis, code operations, version control, and memory/RAG functionality. It supports multiple embedding providers including Ollama (default) and OpenAI-compatible APIs, making it adaptable to different deployment environments and preferences.

Try asking AI

After installing, here are 6 things you can ask your AI assistant:

• get_context_tree

  Get structural AST tree of a project with file headers and symbol ranges.

  Try:Show me the project structureTry:Get the context tree for this codebase
• get_file_skeleton

  Get function signatures, class methods, and type definitions without full bodies.

  Try:Show me the API surface of this fileTry:Get the skeleton of main.py
• semantic_code_search

  Search code by meaning using embeddings over file headers/symbols.

  Try:Find functions related to authenticationTry:Search for code that handles user login
• semantic_identifier_search

  Retrieve functions/classes/variables with ranked call sites.

  Try:Find all usages of calculateTotalTry:Show where processRequest is called
• get_blast_radius

  Trace files and lines where a symbol is imported or used.

  Try:Show impact of changing User classTry:Get blast radius for delete method
• run_static_analysis

  Run linters to find unused variables, dead code, and type errors.

  Try:Check for unused variablesTry:Run static analysis on the backend
• propose_commit

  Write code with validation and create restore points.

  Try:Propose a fix for the bugTry:Update the authentication logic
• get_feature_hub

  Navigate Obsidian-style feature hubs with wikilinks.

  Try:Show the feature hub for paymentsTry:Navigate the documentation hub
• search_memory_graph

  Semantic search with graph traversal for direct matches and neighbors.

  Try:Search for authentication patternsTry:Find related concepts to caching
• upsert_memory_node

  Create or update memory nodes with auto-generated embeddings.

  Try:Remember this API endpointTry:Add a note about the payment flow
• create_relation

  Create typed edges between memory nodes.

  Try:Link User to AccountTry:Relate Order to Payment
• undo_change

  Restore files to state before specific AI changes.

  Try:Undo the last changeTry:Restore to before payment feature

Comparable tools

Installation

{
  "mcpServers": {
    "contextplus": {
      "command": "bunx",
      "args": ["contextplus"],
      "env": {
        "OLLAMA_EMBED_MODEL": "nomic-embed-text",
        "OLLAMA_CHAT_MODEL": "gemma2:27b",
        "OLLAMA_API_KEY": "YOUR_OLLAMA_API_KEY"
      }
    }
  }
}

{
  "servers": {
    "contextplus": {
      "type": "stdio",
      "command": "bunx",
      "args": ["contextplus"],
      "env": {
        "OLLAMA_EMBED_MODEL": "nomic-embed-text",
        "OLLAMA_CHAT_MODEL": "gemma2:27b",
        "OLLAMA_API_KEY": "YOUR_OLLAMA_API_KEY"
      }
    }
  },
  "inputs": []
}

FAQ

Which programming languages does Context+ support?

Context+ supports multiple languages through Tree-sitter with 43 extensions, including TypeScript, Python, Rust, Go, and others.

Can I use Context+ with different embedding models?

Yes, Context+ supports both Ollama (default) and OpenAI-compatible APIs, including Gemini, OpenAI, Groq, and vLLM.

Compare contextplus with