Turn your project into a visual map β you and your AI assistant can ask questions and get instant answers, no code hunting required
Ever opened a new repo and felt lost β unsure which files matter, which modules depend on what, or where data flows from and to?
That's exactly why Graphify v8 exists. It transforms your codebase into a knowledge graph β a living map of your project β so you and your AI assistant can query, explore, and understand it many times faster.
Imagine you just got handed a new project β hundreds of files, no idea where to start, which files matter, or what depends on what?
Graphify is like a project scanner β it reads all the code and creates a map that helps you:
| Before (without Graphify) | After (with Graphify) |
|---|---|
| Open a new project β manually hunt through files | Run one command β instantly get a full project map |
| Ask AI "which file handles login?" β AI must re-read all the code | Ask AI β AI "looks" at the existing map β answers right away |
| Hard to know which file depends on which | The map clearly shows how files are connected |
| Every new AI session needs the project "explained" from scratch | The map is saved β any AI session understands your project instantly |
Requires Python 3.10+. The PyPI package name is graphifyy (double y), but the CLI command remains graphify.
# Install
uv tool install graphifyy
# Install skill for Codex
graphify install --platform codex
# Create a map for the current project
$graphify .
$graphify. For deeper analysis, enable multi_agent = true in your Codex config file.
Graphify does more than create maps β it helps you interact with your project in many ways:
Run $graphify . for the current project, or point it to a folder / GitHub link.
$graphify .
$graphify /path/to/project
$graphify https://github.com/owner/repo
Use --mode deep for deeper analysis; --no-viz when you only need data.
Code changed? Refresh the map:
$graphify . --update
You can set it to auto-update on every commit:
graphify hook install
Ask Graphify like you'd ask a person:
$graphify query "Which module depends on the database?"
If a map already exists, Graphify answers instantly β no re-reading code.
See how two files or concepts are connected:
$graphify path "API endpoint" "DatabaseModel"
Summarize the role and connections of a specific part of the project:
$graphify explain "SessionFactory"
Add documentation, PDFs, or external links to the map:
$graphify add https://docs.example.com/api
Export the map in multiple formats to share with your team:
graphify export callflow-html
$graphify . --wiki
graphify merge-graphs a.json b.json
Supports SVG, GraphML, Neo4j, FalkorDB, Obsidian, and wiki.
Commit graphify-out/ so everyone and every AI assistant shares the same map.
Graphify respects .gitignore and supports .graphifyignore to exclude sensitive files.
Let other tools access the map via MCP:
$graphify . --mcp
python -m graphify.serve graphify-out/graph.json
Save results and reflect on accumulated knowledge:
graphify save-result ...
graphify reflect
Everything is saved in the graphify-out/ folder:
| File | What Does It Do? |
|---|---|
graph.html |
An interactive view to explore your project with your mouse |
GRAPH_REPORT.md |
A summary report of the project structure and key points |
graph.json |
The map data β for querying, sharing, or connecting tools |
Every relationship in the map is marked with a trust level:
EXTRACTED β Fact from the code, read directly β most reliableINFERRED β Inferred from context β fairly reliableAMBIGUOUS β Unclear β should be double-checkedGraphify is especially useful when:
Graphify v8 helps you understand projects many times faster β instead of hunting through files one by one, you just run one command and get a full project map.
Especially when used with an AI assistant β instead of "explaining" your project to AI every time, the AI can "look" at the existing map and answer your questions instantly.
Have you tried Graphify yet? Drop a comment if you have questions or want to see a specific demo!