Repolode
Enviar
Graphify-Labs

Graphify-Labs/graphify

Turns a codebase, its docs, schemas and PDFs into a queryable knowledge graph for AI coding assistants, built locally with tree-sitter AST parsing.

Ver no GitHubSite
Estrelas
128,693
Forks
12,401
Issues abertas
1,555
Último commit
há 3 dias

O que é?

Graphify maps a project into a knowledge graph so an AI coding assistant can answer questions by traversing relationships instead of grepping through files. Running /graphify in Claude Code, Cursor, Codex, Gemini CLI and other supported assistants produces an interactive graph.html, a GRAPH_REPORT.md summary of key concepts and communities, and a graph.json that can be queried later with graphify query, path and explain.

Code is parsed locally with tree-sitter AST across roughly 40 languages, without an LLM and without sending code off the machine. Docs, PDFs, images and audio or video go through an optional semantic pass that uses the assistant's model or a configured API key. Each edge is tagged EXTRACTED or INFERRED so readers can tell what was read directly from the source from what was resolved.

The project is a Python package installed with uv or pipx, and it provides a CLI, per-assistant skill installation, optional extras for PDF, SQL, Neo4j and other backends, and an MCP server.

Recursos principais

  • 01Builds a knowledge graph from code, docs, SQL schemas, configs and PDFs.
  • 02Parses about 40 programming languages locally with tree-sitter AST, with no LLM required for code.
  • 03Tags each edge as EXTRACTED or INFERRED to separate explicit links from inferred ones.
  • 04Answers questions with query, shortest-path tracing with path, and concept explanation with explain.
  • 05Installs as a skill for Claude Code, Cursor, Codex, Gemini CLI and other assistants, and offers an MCP server.

Casos de uso

  • Onboarding onto an unfamiliar codebase by asking an assistant how modules connect.
  • Letting AI coding agents look up architecture relationships instead of re-reading many files each session.
  • Tracing dependencies between services, schemas and documentation in a single graph.
  • Sharing a committed graph.json so a team queries the same project context.

Tópicos

#knowledge-graph#code-analysis#ai-agents#tree-sitter#code-search#claude-code#llm#antigravity#ast#codex

Perguntas frequentes

O que é graphify?
Turns a codebase, its docs, schemas and PDFs into a queryable knowledge graph for AI coding assistants, built locally with tree-sitter AST parsing. Graphify maps a project into a knowledge graph so an AI coding assistant can answer questions by traversing relationships instead of grepping through files. Running /graphify in Claude Code, Cursor, Codex, Gemini CLI and other supported assistants produces an interactive graph.html, a GRAPH_REPORT.md summary of key concepts and communities, and a graph.json that can be queried later with graphify query, path and explain.
graphify é gratuito? Qual é a licença?
graphify é open source sob a licença Apache-2.0. Confira os termos da licença antes de usá-lo em produtos comerciais.
graphify ainda é mantido?
Manutenção ativa. O último commit de graphify foi em 16 de out. de 2026.
Para que posso usar graphify?
Onboarding onto an unfamiliar codebase by asking an assistant how modules connect. Letting AI coding agents look up architecture relationships instead of re-reading many files each session. Tracing dependencies between services, schemas and documentation in a single graph. Sharing a committed graph.json so a team queries the same project context.
Onde consigo graphify?
O código-fonte e as instruções de instalação estão no GitHub: https://github.com/Graphify-Labs/graphify

Resumo gerado por IA a partir do README do projeto. As estatísticas do GitHub são atualizadas diariamente.

Projetos relacionados