graphify-rs

A Rust rewrite of graphify — an AI-powered knowledge graph builder that transforms code, docs, papers, and images into queryable, interactive knowledge graphs.

中文文档

Compared to the Python Version

Full feature parity with the original, plus:

Output format is fully compatible — graph.json uses the same NetworkX node_link_data schema, so Python tools can read Rust output and vice versa.

Installation

From crates.io

cargo install graphify-rs

From source

git clone https://github.com/TtTRz/graphify-rs.git
cd graphify-rs
cargo install --path .

Quick Start

graphify-rs build
open graphify-out/graph.html

CLI Usage

Build

graphify-rs build                               # build current directory
graphify-rs build --path . --output graphify-out
graphify-rs build --format json,html,report      # select export formats
graphify-rs build --code-only                    # skip docs/papers
graphify-rs build --update                       # incremental rebuild
graphify-rs build --no-llm                       # skip Claude API

Query & Analyze

graphify-rs query "how does authentication work?"
graphify-rs query "error handling" --dfs --budget 3000
graphify-rs diff old/graph.json new/graph.json
graphify-rs stats graphify-out/graph.json

Watch & Serve

graphify-rs watch --path . --output graphify-out  # auto-rebuild on changes
graphify-rs serve --graph graphify-out/graph.json  # start MCP server
graphify-rs ingest https://arxiv.org/abs/2301.00001

Platform Integrations & Hooks

graphify-rs claude install    # Claude Code
graphify-rs codex install     # Codex
graphify-rs hook install      # Git pre-commit hook

Global Flags

graphify-rs -q build          # quiet mode
graphify-rs -v build          # verbose/debug mode
graphify-rs -j 4 build        # limit parallel jobs

Utilities

graphify-rs completions bash > ~/.bash_completion.d/graphify-rs
graphify-rs completions zsh > ~/.zfunc/_graphify-rs
graphify-rs init              # creates graphify.toml

Agent Usage (Skill)

graphify-rs can be used as an AI coding agent skill. After installing the integration, your agent (Claude Code, Codex, etc.) automatically gets access to the knowledge graph.

Setup

# Install skill for your platform
graphify-rs claude install    # writes to .claude/settings.json + CLAUDE.md
graphify-rs codex install     # writes to .codex/hooks.json

# Build the graph (agent can also do this)
graphify-rs build

How Agents Use It

Once installed, the agent will:

1. Before answering architecture questions — check if graphify-out/graph.json exists
2. If it exists — read graphify-out/GRAPH_REPORT.md for project overview
3. For specific questions — run graphify-rs query "<question>" to get relevant subgraph context
4. For ongoing work — the MCP server (graphify-rs serve) provides 7 tools the agent can call directly

MCP Server Integration

Add to your Claude Code MCP config:

{
  "mcpServers": {
    "graphify": {
      "command": "graphify-rs",
      "args": ["serve", "--graph", "graphify-out/graph.json"]
    }
  }
}

The agent can then call tools like query_graph, get_node, get_neighbors, god_nodes, etc. directly.

Configuration

Create graphify.toml in your project root (or run graphify-rs init):

output = "graphify-out"
no_llm = false
code_only = false
formats = ["json", "html", "report"]

CLI flags always override config file values.

Architecture

14 crates organized as a Cargo workspace:

Output Formats

MCP Server Tools

When running graphify-rs serve, 7 tools are available over JSON-RPC 2.0 (stdio):

Supported Languages (21)

Contributing

See CONTRIBUTING.md for development setup, code style, and PR guidelines.

License

MIT — see LICENSE.

This project is a Rust rewrite of graphify by safishamsi.