Skip to content

Shahrainis/ai-git-bot

Β 
Β 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

454 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

AI-Git-Bot

License: MIT Docker Pulls GitHub release GitHub stars GitHub issues

🌐 Languages: English Β· δΈ­ζ–‡ Β· ν•œκ΅­μ–΄ Β· ζ—₯本θͺž

Automate the necessary-but-uncomfortable parts of software development β€” directly inside the Git tools your team already uses.

Every team has a list of "we know we should be doing this" engineering chores: writing a properly scoped issue before coding starts, adding a regression test for the bug you just fixed, re-reviewing a PR after the third force-push, tearing down a stale preview environment. These chores are necessary but uncomfortable β€” and they get cut first under deadline pressure.

AI-Git-Bot turns those chores into repeatable, automated workflows that live natively inside Gitea, GitHub, GitHub Enterprise, GitLab, and Bitbucket Cloud β€” triggered by the events your team is already producing (issue assigned, PR opened, reviewer re-requested, @bot mentioned in a comment). Self-hostable end-to-end, including local LLMs β€” nothing has to leave your infrastructure.

πŸ“£ New here? Read the pitch β€” why this project exists, what it does for your team, and how it compares to Copilot Workspace / GitLab Duo / Qodo / Aider (~10 min read).

AI-Git-Bot Architecture Schema

✨ What it does

Workflow Triggered by What it produces
PR Review PR opened with bot as reviewer, or review re-requested Inline + summary review comments, chunked for large diffs
Interactive Q&A @bot mention in any PR or inline review comment Threaded reply with file/diff context and session memory
Issue β†’ Code (coding agent) Issue assigned to a coding bot A pull request implementing the change, validated with your project's own build tooling
Issue β†’ Better Issue (writer agent) Issue assigned to a writer bot A structured AI Created Issue with acceptance criteria
Unit tests (test author) PR opened, or @bot generate-tests White-box unit tests for the diff, run with your project's test runner and committed onto the PR branch
Full-stack QA (E2E tests) PR opened on a bot with a deployment target A generated Playwright suite run against a per-PR preview environment, report posted to the PR, teardown on close

All workflows are opt-in per bot β€” pick the chore that hurts most, wire one bot, done. Nothing changes for repos you don't touch.

πŸŽ₯ Watch the PR workflows in action: AI-Git-Bot β€” PR workflow walkthrough on YouTube

AI-Git-Bot Dashboard

πŸ“Έ Screenshots: reviews, conversations, and coding agents across platforms

Gitea: Gitea Code Review

GitHub: GitHub Code Review

GitLab: GitLab Code Review

Bitbucket: Bitbucket Code Review

Coding agent (GitHub): GitHub Agent

πŸ”Œ Mix & match any AI provider with any Git platform

AI-Git-Bot is a small, self-hosted gateway: configure an AI provider once, attach it to as many bots and repositories as you like. API keys, prompts, and tool whitelists are managed in one admin UI; secrets are encrypted at rest (AES-256-GCM); remote MCP servers can be attached with a per-tool whitelist.

AI providers Git platforms
Anthropic (Claude) Gitea (self-hosted)
OpenAI (+ OpenAI-compatible APIs) GitHub / GitHub Enterprise
Google AI / Gemini GitLab (gitlab.com & self-managed)
Ollama (local LLMs) Bitbucket Cloud
llama.cpp (local GGUF models)

πŸ§ͺ Project maturity: Gitea and GitHub are well-tested in production use; GitLab and Bitbucket Cloud are experimental (implemented from the official API docs and smoke-tested). The Full-stack QA / E2E workflow is the most complex moving part and should be considered experimental on every provider. Bug reports are very welcome β€” every non-trivial workflow ships with a reproducible docker-compose system-test stack; see the Testing Guide.

πŸš€ Quick Start

Run with Docker Compose (one app container + PostgreSQL β€” no Kubernetes required):

git clone https://github.com/tmseidel/ai-git-bot.git
cd ai-git-bot
docker compose up --build -d

Then:

  1. Open http://localhost:8080 and create your administrator account
  2. Create an AI Integration (provider + API key)
  3. Create a Git Integration (Gitea Β· GitHub Β· GitLab Β· Bitbucket)
  4. Create a Bot, enable the workflows you want, and copy its Webhook URL
  5. Configure the webhook in your Git provider β€” done!

➑️ Full instructions: Deployment and the Admin Guide. The image is on Docker Hub.

πŸ“š Documentation

The documentation is organized by audience in the Documentation Hub:

You are a… Start here
πŸ‘€ User β€” a bot is already set up, you just use the Git platform Using the Bot
πŸ› οΈ Administrator β€” you set up the software, bots, and workflows Deployment Β· Admin Guide
πŸ§ͺ Tester β€” you want to try out features safely Testing Guide
πŸ’» Developer β€” you work with the code Local Development Β· Architecture

License

MIT

About

AI-Git-Bot is a lightweight, self-hostable gateway application that connects your Git platforms with AI providers

Resources

License

Code of conduct

Contributing

Security policy

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages

  • Java 90.9%
  • HTML 8.0%
  • Other 1.1%