Tangkapan layar proyek
Ringkasan
Graphify transforms software projects into interactive knowledge graphs that can be queried instead of searched. By utilizing tree-sitter for deterministic AST parsing, it maps code structures, cross-file dependencies, and documentation into a graph format. It distinguishes between explicitly extracted relationships and inferred ones, providing transparency into how connections are made. The tool operates locally for code analysis, ensuring no source code leaves the developer's machine. It integrates directly into AI coding assistants via a skill system, allowing the assistant to query the graph to answer questions, trace paths between components, and explain concepts without reading raw files.
Fitur utama
- Local deterministic AST parsing via tree-sitter
- No vector store or embeddings required
- Confidence tags for every edge (EXTRACTED vs. INFERRED)
- Interactive graph.html for visual exploration
- Cross-file link resolution across ~40 languages
- Leiden community detection with LLM-free labels
- Integration with 20+ AI coding assistants
- Support for docs, PDFs, images, and video/audio mapping
- Strict mode for AI assistants to enforce graph-first queries
Persyaratan, instalasi, dan mulai cepat
Penggunaan
Kompatibilitas model dan kasus penggunaan
Graphify integrates with AI coding assistants including Claude Code, Cursor, Codex, Gemini CLI, GitHub Copilot, Aider, and 15+ others. For semantic passes over docs and media, it can use the assistant's model or configured API keys for OpenAI, Anthropic, Google Gemini, AWS Bedrock, Azure OpenAI, or local Ollama models.
Catatan lisensi dan risiko
Apache-2.0
Editorial verification 2026-08-09: repository URL, owner, description, license and repository statistics were reviewed. License metadata: Apache-2.0. README was fetched for the channel draft; re-check repository dependencies, releases and model terms before production use.
Rilis dan pemeliharaan
Not stated in the repository metadata