DEVOURED
Graphify (GitHub Repo)
Graphify maps codebases and documentation into queryable knowledge graphs locally, bypassing vector search for deterministic codebase navigation.
What: Graphify uses tree-sitter AST parsing to map code in 40+ languages into a knowledge graph. It integrates with 20+ AI coding assistants (e.g., Claude Code, Cursor) and allows queries to trace relationships between code, docs, PDFs, and media without using embeddings.
Why it matters: This signals a trend away from pure semantic search (RAG) toward deterministic, structure-aware code navigation, which is often more accurate for complex architectural queries.
Takeaway: Install via `uv tool install graphifyy` and run `graphify install` to register it with your IDE's AI coding assistant.
Deep dive
- Uses AST parsing (tree-sitter) for 100% deterministic code relationships (imports, calls, inheritance).
- Maps docs, PDFs, and video via an AI semantic pass (or locally with faster-whisper/OCR).
- Exports to graph.json and a browser-viewable interactive graph.html.
- Categorizes edges as 'EXTRACTED' (code-based) or 'INFERRED' (LLM-based) to maintain transparency.
- Provides a CLI for querying relationships and explaining concepts.
- Supports persistent git hooks for incremental graph updates.
- Offers an MCP server mode for repeated, context-aware AI tool access.
Decoder
- AST (Abstract Syntax Tree): A tree representation of the abstract syntactic structure of source code, allowing for precise semantic analysis without executing the code.
- RAG (Retrieval-Augmented Generation): A technique where an AI model retrieves external data from a vector store to improve its responses.
- MCP (Model Context Protocol): An open standard for connecting AI assistants to data sources and tools consistently.
Original article
Full article content is not available for inline reading.
Explore all 53 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi