Tired of manually crafting Git commit messages? Supercharge your Python development workflow with avcmt-py, the intelligent CLI tool that automates semantic, structured commit message generation using AI. Boost productivity and maintain a crystal-clear repository history effortlessly.
TL;DR: AI-powered, semantic-release-style git commit automation for Python projects. One command, zero guesswork: meaningful, structured commits using your Pollinations AI API Token. Install, configure your API key, and enjoy never writing boring commit messages again!
AVCMT-PY (Automated Version Control & Management Tool for Python) is a blazing-fast, fully-automated CLI tool for generating meaningful, structured git commit messages using AI (Gemini, OpenAI, Pollinations, etc) — optimized for Python developers who want clean semantic-release workflow, better productivity, and crystal-clear repo history.
- No more generic "fix bug", "update code" commits.
- Just run
avcmtand get a ready-to-commit, semantic-release-formatted message, automatically grouped by directory or file. - Integrates with pre-commit, CI/CD, and release workflows.
- Flexible AI provider: choose your favorite (support for Gemini, Pollinations, OpenAI out-of-the-box).
-
AI-Powered Commit Messages: Generate detailed, semantic-release-style commit messages from git diff with a single command.
-
Directory Grouping: Automatically groups and commits related changes per directory (or as a catch-all).
-
Semantic Release Ready: Commit format fully compatible with semantic-release for auto versioning & changelogs.
-
Multi-Provider AI: Easily switch between Gemini, Pollinations, OpenAI (or extend to your own LLM API).
-
Jinja2 Prompt Templates: Fully customizable prompt rendering using Jinja2 templates for flexible commit messaging.
-
Zero Hardcoded Secrets: API keys are loaded from
.envor environment variables. -
Easy to Install, Easy to Use: Works on any Python project, no lock-in.
-
Developer Tools Included: Scripts for linting, formatting, preflight check, and triggering semantic release.
-
Pre-commit & CI/CD Friendly: Fully integrated with pre-commit and GitHub Actions for automated workflows.
pip install avcmt-pyOr install from source:
git clone https://github.com/andyvandaric/avcmt-py.git
cd avcmt-py
pip install .-
Add your API key
-
Copy
.env.exampleto.env -
Edit
.envand fill your Pollinations or OpenAI API key
# Example (.env) OPENAI_API_KEY=your_openai_token POLLINATIONS_TOKEN=your_pollinations_token
-
-
(Optional) Enable pre-commit hook:
pre-commit install -
Run avcmt:
avcmt # AI generates & applies grouped commits!-
Optionally use:
-
--dry-run(preview messages) -
--push(auto-push after commit) -
--debug(show AI prompts & raw response)
-
-
-
Done!
- Check your git log for clean, structured, semantic-release-ready commit messages.
avcmt [OPTIONS]-
--dry-run: Preview commit messages without applying -
--push: Push commits to remote after done -
--debug: Show debug info (prompts & AI response)
avcmt --dry-run
avcmt --push-
Place
.envin your project root (or set env vars globally) -
Supported ENV:
-
POLLINATIONS_API_TOKEN -
(other providers: applied next update)
-
avcmt-py/
├── .env.example
├── .github/
│ └── workflows/
│ ├── pre-commit.yml
│ └── release.yaml
├── .gitignore
├── .pre-commit-config.yaml
├── CHANGELOG.md
├── LICENSE
├── README.md
├── pyproject.toml
├── avcmt/
│ ├── __init__.py
│ ├── ai.py
│ ├── cli.py
│ ├── commit.py
│ ├── utils.py
│ ├── prompt_templates/
│ │ └── commit_message.j2
│ └── providers/
│ ├── __init__.py
│ ├── openai.py
│ └── pollinations.py
├── scripts/
│ ├── check.py
│ ├── clean.py
│ ├── format.py
│ ├── helper.py
│ ├── lintfix.py
│ ├── preflight.py
│ ├── semrel.py
│ └── setup.py
└── setup.cfg-
8000
-
avcmt/cli.py--- CLI entry point, handles argument parsing and triggers auto-commit. -
avcmt/commit.py--- Core logic for grouping, git interaction, and structured AI commit message generation. -
avcmt/ai.py--- Manages prompt rendering and request to the AI provider (Jinja2-powered). -
avcmt/utils.py--- Helper functions for environment, logging, and file operations. -
avcmt/__init__.py--- Marks the core package. -
avcmt/prompt_templates/commit_message.j2--- Jinja2 template for AI commit prompt. -
avcmt/providers/openai.py--- Adapter for OpenAI API. -
avcmt/providers/pollinations.py--- Adapter for Pollinations API. -
avcmt/providers/__init__.py--- Provider interface loader. -
scripts/check.py--- Run validation checks on repo status. -
scripts/clean.py--- Optional cleanup utility. -
scripts/format.py--- Format code using Ruff or Black. -
scripts/helper.py--- Shared utilities across scripts. -
scripts/lintfix.py--- Lint and auto-fix with Ruff. -
scripts/preflight.py--- Pre-commit safety check runner. -
scripts/semrel.py--- Trigger python-semantic-release publish process. -
scripts/setup.py--- One-shot setup script for dev environment. -
.env.example--- Environment file template. Copy to.envand fill your token. -
.pre-commit-config.yaml--- Pre-commit hook configuration. -
.gitignore--- Ignore compiled files, .env, cache, etc. -
pyproject.toml--- Project metadata and dependency configuration. -
setup.cfg--- Optional setup file for tools compatibility. -
README.md--- Full project description and usage. -
LICENSE--- MIT License. -
CHANGELOG.md--- Auto-generated changelog via semantic release. -
.github/workflows/release.yaml--- CI workflow for auto versioning and publishing. -
.github/workflows/pre-commit.yaml--- CI pre-commit hook runner.
-
Custom AI Providers: See
avcmt/ai.pyto extend with your own LLM API. -
Integration with pre-commit: Works out-of-the-box, can be called in hooks or CI.
-
Full CLI options: Run
avcmt --helpfor all flags.
Q: Will this overwrite my changes or commit everything automatically? A: No, only staged files are affected. You're always in control.
Q: Can I use it for monorepos? A: Yes, directory grouping is automatic, but fully configurable.
Q: What if my provider's API token is missing or invalid? A: You'll see a clear error and nothing will be committed.
Q: Is it safe for public/private repos? A: Yes, no token or diff is ever sent to any server except the AI you choose.
-
✨ Stop wasting time on commit messages
-
✨ Zero learning curve, drop-in to any Python repo
-
✨ Works everywhere: CLI, hook, CI/CD, local/dev/remote
-
✨ Your AI, your rules: bring your own API key, use any LLM
Apache 2.0 Made by Andy Vandaric
-
Inspired by semantic-release, and real-life productivity pain points.
-
Powered by Pollinations AI.