AICommit
The Problem Every Developer Knows
We've all been there:
git commit -m "fix"
git commit -m "update stuff"
git commit -m "changes"
git commit -m "wip"
git commit -m "final fix"
git commit -m "final fix for real"The result? A git history that's impossible to navigate, confused teammates, and failed code reviews.
The Solution: AI-Powered Commits
AICommit analyzes your staged changes and generates professional commit messages that follow the Conventional Commits specification:
git add .
aicommit
# Output:
# ✨ feat(auth): implement OAuth2 social login
#
# Add JWT-based authentication system with password hashing.
# - Create login/register endpoints with validation
# - Implement secure password hashing with bcrypt
# - Add JWT token generation and verification middlewareWhat Makes AICommit Different
100% Private and Local
Unlike other AI commit tools that send your code to cloud APIs, AICommit runs entirely on your machine using LM Studio. Your code never leaves your computer.
- No API keys to manage
- No subscription fees
- No data collection
- Works completely offline
Context-Aware Messages
AICommit adapts the commit message length based on the scope of your changes:
Small change (1-2 files):
🐛 fix(ui): correct button alignment issueMedium change (3-5 files):
✨ feat(auth): implement user registration
Add complete user registration system with validation.
- Create registration form with input validation
- Add email confirmation workflow
- Implement password strength requirementsLarge refactor (6+ files):
♻️ refactor(api): restructure authentication system
Modernize authentication architecture for better security.
- Replace JWT with session-based authentication
- Add OAuth2 integration for social login
- Implement role-based access control (RBAC)
- Create authentication middleware pipeline
- Add comprehensive security headersQuick Start
Installation
npm install -g @abd3lraouf/aicommitSetup Local AI
- Download LM Studio
- Search and download:
qwen3-4b-teen-emo(4GB, fast & accurate) - Load the model and start the server
Configure
aicommit configUse It
git add .
aicommitCommand Options
# Basic usage
aicommit
# Interactive mode - review before committing
aicommit --interactive
# Dry run - see message without committing
aicommit --dry-run
# Verbose output
aicommit --verbose
# Debug mode
aicommit --debugCommit Types
AICommit automatically selects the appropriate type and emoji:
| Type | Emoji | Usage |
|---|---|---|
| feat | ✨ | New features |
| fix | 🐛 | Bug fixes |
| docs | 📝 | Documentation |
| refactor | ♻️ | Code restructuring |
| perf | ⚡ | Performance improvements |
| test | ✅ | Adding tests |
| chore | 🔧 | Maintenance tasks |
Technical Details
Built with TypeScript and designed for extensibility:
- Analyzes git diff to understand changes
- Constructs intelligent prompts for the AI
- Parses and formats AI responses
- Supports multiple LLM providers via OpenAI-compatible API