Frequently Asked Questions

Everything you need to know about Graphify, the AI-powered code assistant that turns your codebase into a queryable knowledge graph.

Get Started View on GitHub

Common Questions

Find answers to the most frequently asked questions about Graphify

What is Graphify? ▼

Graphify is an AI coding assistant that maps your entire project into a queryable knowledge graph. It processes code, documentation, SQL schemas, R scripts, shell scripts, papers, images, and videos into a single interconnected graph that you can query instead of grepping through files. It works with major AI assistants like Claude Code, Cursor, Codex, Gemini CLI, and GitHub Copilot.

How does Graphify work? ▼

Graphify uses a three-step process:

  1. Code Analysis: Uses tree-sitter AST to parse code locally (no LLM, nothing leaves your machine) and extract relationships between functions, classes, and modules.
  2. Knowledge Graph Construction: Creates nodes for every code entity and documentation item, with edges showing relationships like imports, calls, and inheritance.
  3. Query Interface: Provides natural language queries, path tracing, and explanations to help you understand your codebase instantly.

Every relationship is tagged as EXTRACTED (explicit in source) or INFERRED (derived by resolution), so you always know what was found versus guessed.

What programming languages does Graphify support? ▼

Graphify supports 36+ programming languages through tree-sitter grammars, including:

  • Python, TypeScript, JavaScript, Go, Rust, Java, C/C++, Ruby, C#, Kotlin, Swift, and more
  • Specialized support for Salesforce Apex, Terraform/HCL, and SQL schemas
  • Package manifests like pyproject.toml, go.mod, pom.xml
  • Documentation in Markdown, HTML, YAML, and text formats

Full list available in the GitHub repository.

Is Graphify free to use? ▼

Yes! Graphify is open-source under the MIT License. The core functionality for code mapping is completely free and works locally without any API calls. For documentation and media processing, you can use local models (Ollama) or your preferred AI assistant's API. The project is funded by GitHub Sponsors and community contributions.

What file types can Graphify process? ▼

Graphify handles a wide range of file types:

  • Code: 36+ programming languages via tree-sitter AST
  • Documentation: Markdown, HTML, YAML, text files
  • Office: Word documents (.docx), Excel spreadsheets (.xlsx)
  • PDFs: Research papers and documentation
  • Media: Images (PNG, JPG, etc.), videos (MP4, MOV), audio files
  • Online: YouTube videos and web pages

Video and audio transcription uses local faster-whisper for privacy.

How do I install Graphify? ▼

Installation takes just 30 seconds:

  1. Install the CLI: uv tool install graphifyy (or pipx install graphifyy)
  2. Register with your AI assistant: graphify install
  3. Build your knowledge graph: /graphify . in your AI assistant

This creates three files:

  • graph.html - Interactive visualization
  • GRAPH_REPORT.md - Key insights & connections
  • graph.json - Full graph data

What are the outputs of Graphify? ▼

Graphify generates three key outputs:

  1. graph.html: An interactive, clickable visualization of your codebase as a force-directed graph. You can explore communities, filter nodes, and search concepts.
  2. GRAPH_REPORT.md: A comprehensive report highlighting:
    • Most-connected concepts ("God nodes")
    • Surprising connections across modules
    • Design rationale from code comments
    • Suggested questions only the graph can answer
  3. graph.json: The full knowledge graph in JSON format for programmatic access.
How do I query the knowledge graph? ▼

Use natural language queries directly in your AI assistant:

  • graphify query "what connects auth to the database?"
  • graphify path "UserService" "DatabasePool" - Find shortest path
  • graphify explain "RateLimiter" - Get detailed explanation

These commands return precise answers about your codebase without reading files, leveraging the knowledge graph structure.

Can I use Graphify with my favorite AI coding assistant? ▼

Yes! Graphify integrates with 20+ AI coding assistants:

  • Claude Code, Cursor, Codex, Gemini CLI, GitHub Copilot
  • OpenCode, Kilo Code, Aider, OpenClaw, Factory Droid
  • Trae, Hermes, Kimi Code, Pi, Devin CLI
  • And more with the generic Agent Skills framework

Each platform has a specific install command like graphify claude install or graphify cursor install. See the full list in the documentation.

How does Graphify handle privacy? ▼

Graphify prioritizes privacy with a local-first approach:

  • Code parsing: Uses tree-sitter AST locally with no API calls
  • Video/audio: Transcribed locally with faster-whisper
  • Documentation: Sent to your AI assistant's model API
  • Local options: Use Ollama for LLM processing without external APIs
  • No telemetry: No usage tracking or analytics

For sensitive projects, you can run everything locally with --backend ollama.

What are the system requirements? ▼

Graphify runs on major platforms with minimal requirements:

  • OS: Windows, macOS, Linux
  • Python: 3.10+ (included in installers)
  • Memory: 4GB RAM minimum, 8GB+ recommended
  • Storage: 1GB for installation + project files
  • Optional: GPU for faster video transcription

Quick install commands provided for Homebrew, winget, apt, and more.

How do I contribute to Graphify? ▼

We welcome community contributions! Ways to contribute:

  • Code contributions: Submit pull requests for new language support or features
  • Documentation: Help improve the docs and README
  • Testing: Test on your codebase and report issues
  • Sharing examples: Share your use cases with worked examples
  • Sponsoring: Sponsor via GitHub Sponsors

See the GitHub repository for contribution guidelines and development setup.