RAG Frameworks · Graphify-Labs

graphify

Graphify turns your codebase, docs, PDFs, images, and videos into a queryable knowledge graph that you can explore instead of grep through files. It works as a skill in Claude Code, Cursor, and other AI assistants, parsing code locally with tree-sitter (no LLM) and optionally enriching with semantic analysis for non-code content.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/Graphify-Labs/graphify

79.8k

GitHub stars

7.9k

Forks

Python

Primary language

MIT

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	Graphify-Labs/graphify
Owner	Graphify-Labs
Primary language	Python
License	MIT — OSI-approved
Stars	79.8k
Forks	7.9k
Open issues	440
Latest release	v0.9.9 (2026-07-07)
Last updated	2026-07-08
Source	https://github.com/Graphify-Labs/graphify

What graphify is

A Python tool that builds a knowledge graph using tree-sitter AST parsing for code (deterministic, local-first) and optional LLM-based semantic extraction for docs/media. Outputs an interactive HTML graph, JSON representation, and CLI query/path/explain commands. Edges are tagged EXTRACTED (explicit) or INFERRED (resolved), enabling distinction between direct reads and derived connections.

Quickstart

Get the graphify source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/Graphify-Labs/graphify.git2cd graphify3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Rapid onboarding to large codebases

New engineers can query the graph instead of reading scattered files; ask 'what calls X' or 'trace path from module A to module B' without grepping.

Architecture discovery and documentation

Automatically identifies communities (subsystems) and god nodes (high-connectivity concepts), revealing structure that may be missing or outdated in docs.

AI-assisted code understanding in IDEs

Integrates with Claude Code, Cursor, and other assistants as a `/graphify` skill, letting AI reason over the full project graph instead of file-by-file snippets.

Implementation considerations

Requires Python 3.10+; install via uv (recommended) or pipx. Quick setup: `uv tool install graphifyy && graphify install`.
Register the skill with your target AI assistant (Claude Code, Cursor, Codex, Gemini CLI, GitHub Copilot, or 15+ others); integration method not detailed in excerpt but documented elsewhere.
For semantic enrichment of docs/media, configure an API key (OpenAI, Claude, Gemini, or self-hosted); code-only graphs require no external calls.
Graph size scales with codebase size; no guidance on performance for very large repos (millions of lines) provided.
Output three files: `graph.html` (interactive browser visualization), `GRAPH_REPORT.md` (highlights and suggestions), `graph.json` (queryable snapshot).

When to avoid it — and what to weigh

You need real-time code change tracking — Graphify builds a static snapshot. If your codebase changes frequently, you must re-run; no incremental updates documented.
Your code is in unsupported languages or proprietary syntax — Relies on tree-sitter, which supports ~40 languages. Custom DSLs or languages without tree-sitter parsers will not be parsed correctly.
You require strict data privacy and cannot use external APIs — Code parsing is local, but semantic enrichment of docs/PDFs/images/video uses external LLM APIs (configurable). No option for 100% air-gapped operation stated.
You need zero false-positive inferred connections — INFERRED edges are derived by resolution logic; no false-positive rate or precision metrics are disclosed. Validation required for critical dependency analysis.

License & commercial use

Licensed under MIT (MIT License). This is a permissive OSI license allowing commercial use, modification, and redistribution without liability.

MIT license permits commercial use without restriction. However, review your use of external LLM APIs (OpenAI, Claude, Gemini, etc.) during semantic enrichment—those services have separate commercial terms. The tool itself imposes no commercial restrictions.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Low
DEV.co fit	Strong
Assessment confidence	High

Security considerations

Code is parsed locally with tree-sitter (no LLM call, no data egress for code parsing). Semantic pass over docs/media and optional model use sends content to external APIs (OpenAI, Claude, Gemini); review your data sensitivity and API provider terms. No intrinsic vulnerability analysis or exploit detection; used for code understanding, not security scanning. No claims about input validation, injection, or dependency security provided in excerpt.

Alternatives to consider

mem0 / Supermemory

Both offer code memory and RAG for AI assistants but use embeddings and vector search rather than explicit graph structure. Graphify's deterministic AST parsing and EXTRACTED/INFERRED tagging offer more transparency on edge provenance.

GitHub Copilot context window + native IDE search

Free, built-in to GitHub and many IDEs, but does not automate cross-file dependency discovery or provide explicit graph structure. Requires manual query refinement.

Semantic code search tools (e.g., Sourcegraph, Krugle)

Offer code indexing and cross-reference lookup but less emphasis on conversational AI integration or graph community detection. Graphify integrates tightly with AI assistants and highlights subsystem structure.

Software development agency

Build on graphify with DEV.co software developers

Install graphify in 30 seconds and explore your project as an interactive knowledge graph. MIT-licensed, free to use, and fully local for code.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

code-graph-rag graphiti llm-graph-builder

Related on DEV.co

Explore the category and the services that help you build with it.

RAG Frameworks AI development API development Custom software Talk to DEV.co

graphify FAQ

Does graphify send my code to the cloud?

No for code parsing; tree-sitter AST is computed locally. Semantic enrichment of docs, PDFs, images, and videos uses external LLM APIs if configured, so review API provider data policies.

Can I query the graph offline after it's built?

Yes. The `graph.json` snapshot is stored locally. Use `graphify query`, `graphify path`, and `graphify explain` commands without re-running the full build or calling external services.

What programming languages are supported?

Tree-sitter supports ~40 languages (Python, JavaScript, Java, Go, Rust, C/C++, SQL, R, shell, and more). Non-tree-sitter languages will not be parsed as code, though docs/media in any language can be semantically indexed.

How do EXTRACTED and INFERRED edges differ?

EXTRACTED edges are explicit in the source (e.g., `import`, `calls`, `inherits`). INFERRED edges are derived by name resolution or type inference. This tagging lets you distinguish high-confidence direct reads from derived connections.

Work with a software development agency

From first prototype to production, DEV.co delivers software development services around tools like graphify. Our software development agency staffs experienced software developers and web developers for custom software development, web development, integrations, and ongoing support across rag frameworks and beyond.

Start mapping your codebase now

Install graphify in 30 seconds and explore your project as an interactive knowledge graph. MIT-licensed, free to use, and fully local for code.

Talk to DEV.co Schedule a Consultation