Gitingest: Transform GitHub Repositories into LLM-Friendly Code Extracts

Introduction

Gitingest is an innovative tool that streamlines the process of extracting code context from Git repositories, making it readily consumable by Large Language Models (LLMs). Its core functionality allows you to transform any Git repository, whether local or remote, into a clean, prompt-friendly text digest. A unique feature is the ability to simply replace "hub" with "ingest" in any GitHub URL to instantly access a structured extract of the codebase.

Key features include:

Easy code context: Generate a text digest from a Git repository URL or a local directory.
Smart Formatting: Output is optimized for LLM prompts, ensuring clarity and relevance.
Comprehensive Statistics: Get insights into file and directory structure, extract size, and token count.
Versatile Access: Available as a CLI tool, a Python package, and convenient browser extensions.

Installation

Gitingest requires Python 3.8 or newer. For private repositories, a GitHub Personal Access Token (PAT) is necessary.

You can install Gitingest via pip:

pip install gitingest

To include server dependencies for self-hosting, use:

pip install gitingest[server]

Alternatively, pipx is recommended for installing Python applications:

pipx install gitingest

Examples

Browser Extension Usage

The Gitingest browser extensions for Chrome, Firefox, and Edge offer a seamless way to get code digests directly from GitHub. Simply navigate to a GitHub repository page and replace "hub" in the URL with "ingest" to view the prompt-friendly extract.

Command Line Usage

The gitingest CLI tool provides powerful options for analyzing codebases.

Basic usage, writing to digest.txt by default:

gitingest /path/to/directory

From a GitHub URL:

gitingest https://github.com/coderamp-labs/gitingest

From a specific subdirectory within a repository:

gitingest https://github.com/coderamp-labs/gitingest/tree/main/src/gitingest/utils

For private repositories, use the --token option or set the GITHUB_TOKEN environment variable:

gitingest https://github.com/username/private-repo --token github_pat_...

Include repository submodules or gitignored files:

gitingest https://github.com/username/repo-with-submodules --include-submodules
gitingest /path/to/directory --include-gitignored

Output to a specific file or STDOUT:

gitingest /path/to/directory --output my_digest.txt
gitingest /path/to/directory --output -

Python Package Usage

Integrate Gitingest directly into your Python applications.

Synchronous usage:

from gitingest import ingest

summary, tree, content = ingest("path/to/directory")
# or from URL
summary, tree, content = ingest("https://github.com/coderamp-labs/gitingest")

Asynchronous usage:

from gitingest import ingest_async
import asyncio

result = asyncio.run(ingest_async("path/to/directory"))

In Jupyter notebooks, you can use await directly:

from gitingest import ingest_async

summary, tree, content = await ingest_async("path/to/directory")

Why Use Gitingest?

Gitingest simplifies the complex task of preparing codebases for AI consumption. By providing a structured, prompt-friendly text digest, it significantly reduces the effort required to feed relevant code context into LLMs. This is invaluable for tasks like code generation, analysis, documentation, and debugging with AI assistance. Its flexibility, offering CLI, Python library, and browser extension access, ensures it fits seamlessly into various developer workflows. Whether you need a quick overview or detailed code insights, Gitingest delivers optimized output for your AI-driven development needs.