Octocode is an intelligent code indexer and semantic search engine designed to build knowledge graphs of codebases, enhancing developer understanding and assistance. It targets developers seeking deeper insights into their code, offering natural language search, relationship mapping, and AI-powered features within a local-first or cloud-integrated architecture. The primary benefit is accelerated code comprehension and navigation through advanced AI and graph-based analysis.

How It Works

Octocode constructs intelligent knowledge graphs from codebases, enabling semantic search via natural language queries and automatic discovery of file relationships (GraphRAG). It employs AI for features like smart commits and code review, supporting a wide array of programming languages through AST parsing and import/export analysis. The system is optimized using LanceDB and offers flexibility with local or cloud-based embedding models, prioritizing a local-first approach for privacy and performance where feasible.

Quick Start & Requirements

• Primary Install: Use the universal install script: curl -fsSL https://raw.githubusercontent.com/Muvon/octocode/master/install.sh | sh. Alternatives include GitHub Releases or cargo install --git https://github.com/Muvon/octocode.
• Prerequisites:
  • Voyage AI API key (VOYAGE_API_KEY) is required for embeddings (200M free tokens/month).
  • OpenRouter API key (OPENROUTER_API_KEY) is optional for LLM features.
• Links:
  • Installation Guide: [Implied by README content]
  • Quick Start: Provided in README.
  • Full Documentation: [Implied by README content, covering Installation, Getting Started, API Keys, Configuration, Commands, Advanced Usage, MCP, LSP, Memory System, Release Management, Architecture, Performance, Contributing]

Highlighted Details

• Semantic Code Search: Supports natural language queries with multi-query capabilities.
• Knowledge Graph (GraphRAG): Automatically discovers and maps relationships between code files.
• Multi-Language Support: Indexes Rust, Python, JavaScript, TypeScript, Go, PHP, C++, Ruby, JSON, Bash, and Markdown.
• AI-Powered Features: Includes smart commits, code review assistance, and a memory system with semantic search.
• Performance & Flexibility: Leverages LanceDB optimization and supports both local and cloud embedding models.

Maintenance & Community

The project is developed by the Muvon team in Hong Kong. Community support is available via GitHub Issues and email at opensource@muvon.io.

Licensing & Compatibility

This project is licensed under the Apache License 2.0. This license is permissive and generally compatible with commercial use and closed-source projects.

Limitations & Caveats

The local-first offline embedding capability using FastEmbed and SentenceTransformer is currently limited to macOS. A Voyage AI API key is essential for core embedding functionality, and AI features may require an optional OpenRouter API key. Sensitive files are excluded via .gitignore respecting, but the privacy benefits of local embeddings are platform-specific.