How to Install Graphify: Complete Setup Guide for the Graphifyy Python Package

Graphify is distributed as the Python package graphifyy and is installed via uv tool install graphifyy, pipx install graphifyy, or pip install graphifyy, followed by the graphify install command to initialize project-specific assets.

This guide explains how to install Graphify from the Graphify-Labs/graphify repository using multiple Python packaging strategies. Whether you are configuring the knowledge-graph generator for local development or automated CI/CD pipelines, these steps ensure the graphifyy package and its CLI entry points are correctly deployed on your system.

Installation Methods for Graphify

The repository supports four distinct installation paths that vary in isolation level and platform reliability. Each method installs the same underlying package but handles Python interpreter resolution and PATH configuration differently.

Install with uv (Recommended)

The uv tool provides the fastest and most reliable installation on modern macOS and Linux systems. It installs graphifyy into an isolated virtual environment that uv manages automatically, eliminating "module-not-found" errors caused by interpreter mismatches.

uv tool install graphifyy

After installation, the graphify binary is available globally without manual PATH modifications.

Install with pipx (Cross-Platform)

pipx works consistently across macOS, Linux, and Windows by creating an isolated environment specifically for the CLI entry point. This method is ideal when you need a system-wide command that does not interfere with other Python projects.

pipx install graphifyy

pipx automatically links the executable to a location already present in your PATH on most systems.

Install with pip (Fallback)

Direct installation via pip is supported but requires caution. On macOS and Windows, this method may resolve to a different Python interpreter than expected, and you might need to manually add the user binary directory (e.g., ~/.local/bin on Linux) to your PATH.

pip install graphifyy

Verify the installation location and ensure your shell can locate the graphify executable before proceeding.

Install in a Virtual Environment (Development)

For reproducible local development or CI/CD pipelines, create a dedicated virtual environment. This approach is required when installing optional extras such as the Model Context Protocol (MCP) support.

python3 -m venv .venv
.venv/bin/pip install "graphifyy[mcp]"

Activate the environment with source .venv/bin/activate (Linux/macOS) or .venv\Scripts\activate (Windows) to access the graphify command.

Post-Installation Setup

Installing the Python package only places the code on disk. You must run the post-installation command to wire Graphify into your specific project.

Initialize Project Assets

Execute graphify install to trigger the core installer logic located in graphify/install.py. This command performs the following actions:

• Creates the graphify-out/ output directory for generated knowledge graphs
• Copies static assets required for rendering
• Generates platform-specific installation fragments via graphify/skill.py

graphify install

Configure Git Hooks

To enable automatic graph regeneration on every commit, install the Git hooks provided in graphify/hooks.py:

graphify hook install

This command places post-commit and post-checkout hooks that trigger the semantic extraction pipelines defined in graphify/llm.py whenever your repository changes.

Key Source Files and Architecture

Understanding the installation pipeline requires familiarity with these critical files in the Graphify-Labs/graphify repository:

• graphify/install.py – Implements the core installer that materializes runtime assets and creates the graphify-out/ folder.
• graphify/hooks.py – Manages Git hook integration, installing and uninstalling the post-commit and post-checkout triggers.
• graphify/skill.py – Generates platform-specific installation fragments for POSIX, PowerShell, and VS Code environments.
• graphify/llm.py – Provides the language model backends (OpenAI, Anthropic, Gemini) used during automatic graph reconstruction.
• pyproject.toml – Declares package metadata, optional extras like graphifyy[office], and build configuration details.

Summary

• Graphify is packaged as graphifyy and published to Python package indexes.
• Preferred method: uv tool install graphifyy for isolated, reliable installation on modern systems.
• Alternative methods: pipx install graphifyy for cross-platform support, or venv + pip for development workflows requiring extras like [mcp].
• Post-installation required: Run graphify install to execute the installer in graphify/install.py and create project directories.
• Git integration: Use graphify hook install to enable automatic graph updates via graphify/hooks.py.

Frequently Asked Questions

What is the correct package name for Graphify?

The package is named graphifyy (with two y's) on PyPI. Use this exact name with uv, pipx, or pip commands. The CLI entry point after installation is simply graphify.

Why does the pip installation method require PATH configuration?

When using pip install graphifyy without a virtual environment, the executable often installs to a user-specific binary directory (such as ~/.local/bin on Linux) that is not automatically included in your shell's PATH on macOS and some Windows configurations. You must manually add this directory to your PATH environment variable to invoke the graphify command.

How do I install Graphify with MCP support?

Install the optional Model Context Protocol (MCP) extras using the bracket notation: pip install "graphifyy[mcp]" or uv pip install "graphifyy[mcp]" within your virtual environment. This installs additional dependencies required for MCP server integration as defined in pyproject.toml.

Where are the Git hooks configured after installation?

Running graphify hook install modifies the .git/hooks/ directory in your current repository. The command copies hook scripts that invoke the semantic extraction logic from graphify/llm.py after each commit or checkout, ensuring your knowledge graph in graphify-out/ stays synchronized with your codebase.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}