Graphify: Let Coding Agents Query Graphs First
techNmak · x · 2026-07-12
This post introduces Graphify, an 82k-star GitHub project. Its core idea is to have coding agents map the codebase into a queryable knowledge graph instead of repeatedly running grep or opening files one by one. The project links functions, classes, files, SQL schemas, infrastructure, docs, and even PDFs/images/videos as nodes and relationships in a graph. Agents can then trace connections, like how authentication links to the database, or the path from UserService to DatabasePool.
The post emphasizes that this is not traditional RAG: it doesn’t rely on embeddings, vector DBs, or LLMs. It uses tree-sitter locally to parse code, turning calls, imports, and inheritance into graph edges. It also tags relationships as EXTRACTED, INFERRED, or AMBIGUOUS to help agents distinguish hard facts from inferences. Graphify can also install hooks/persistent instructions for assistants like Claude Code, Codex, Cursor, Gemini CLI, and Copilot, ensuring agents query the graph before taking action. The graph can be committed to Git, rebuilt automatically with commits, shared across teams, and exposed via MCP.
Related event: Graphify Maps Codebases into Knowledge Graphs for AI Agents(3 posts)→
More from coding & agent
- HeyGen adds a media-sourcing skill for coding agents with 75k images and 10k tracks — HeyGen · 2026-07-22
- Agent search bottlenecks are now about variance, not raw latency — rohanpaul_ai · 2026-07-22
- LangSmith adds tracing for Pipecat, LiveKit, OpenAI Realtime, and Gemini Live — LangChain · 2026-07-22
- An MCP server signs every AI agent tool call into a verifiable Merkle chain — Funky_Chicken_22 · 2026-07-22
- Annotated transcript of a Claude Code team interview is now available — trq212 · 2026-07-22
- Claude Code skill uses 10 Markdown rules to make outputs ADHD-friendly — alex_verem · 2026-07-22