🤖 aicommit

AI commit messages from your local LLM. Reads git diff --staged, asks Ollama for a clean Conventional Commit, commits.

Everything runs offline. No API keys, no data leaves the box.

---

✨ Features

• 📝 Generates Conventional Commit messages from staged changes
• 🤖 Powered by a local LLM via Ollama (or llama-cpp-python)
• ✍️ Opens the proposal in $EDITOR for a quick approval / edit / regenerate loop
• 📓 changelog subcommand: summarizes a tag range into a CHANGELOG.md entry
• 🔌 Installs as a git subcommand — call it as git aicommit

---

🚀 Quick start

# 1. install a model in Ollama (any chat-instruct model works)
ollama pull qwen2.5-coder:7b

# 2. install aicommit straight from GitHub
pip install git+https://github.com/yumiaura/AICommit

# 3. stage and let it write the message
git add -A
git aicommit

A demo asciicast lives in docs/demo.cast — open it with asciinema play docs/demo.cast or paste into asciinema.org.

Sample interaction:

─── proposed message ────────────────────────────────
fix(parser): handle empty input gracefully

- return [] instead of raising on whitespace-only files
- add regression test for the empty-file case
─────────────────────────────────────────────────────
[ Enter = commit · e = edit · r = regenerate · q = quit ]

Changelog mode:

git aicommit changelog v0.4.0..HEAD --out CHANGELOG.md

---

⚙️ Configuration

Quickest way to get a config file:

aicommit config

That creates ~/.config/aicommit/config.toml from the default template (if it doesn't exist yet) and opens it in $EDITOR — falling back to $VISUAL, then nano. Re-running just re-opens the file; your existing values are never clobbered.

The file looks like this:

[llm]
backend = "ollama"           # ollama | llama-cpp
model   = "qwen2.5-coder:7b"
url     = "http://localhost:11434"
temperature = 0.2

[commit]
style          = "conventional"   # conventional | plain
max_subject_len = 72
include_body    = true

Environment variables override the file (AICOMMIT_MODEL=...), and CLI flags override env. Per-repo overrides go in <repo>/.aicommit.toml.

Useful flags:

aicommit [--model M] [--url URL] [--temperature T] [--style {conventional,plain}]
         [--no-body] [--review] [--review-only] [--print] [-y] [--debug]
aicommit changelog <rev-range> [--out CHANGELOG.md]

---

📦 Stack

• ollama HTTP API (or llama-cpp-python for in-process inference)
• subprocess to call git
• argparse + rich for the CLI
• tomllib (Python 3.11+) for config

---

📄 License

MIT (planned).

Author: @yumiaura