Skip to content

README.md

Latest commit

 

History

History
144 lines (108 loc) · 7.87 KB

README.md

File metadata and controls

144 lines (108 loc) · 7.87 KB
Raw

smart-commit-rs

crates.io CI docs license

cgen generates git commit messages from your staged diff using an LLM, then lets you review, edit, or regenerate the message before anything is committed. It ships as a single ~3 MB executable with no language runtime to install. The executable is cgen; the crates.io package is auto-commit-rs.

Features

  • 13 built-in providers — Groq (default), OpenAI, Anthropic, Gemini, Grok, DeepSeek, OpenRouter, Mistral, Together, Fireworks, Perplexity, LM Studio, Ollama — plus any OpenAI-compatible custom endpoint. Local models need no API key.
  • Review before committing — accept, regenerate, edit in your editor, or cancel.
  • Flexible generation — stage tracked changes with --all, print a clean message with --stdout, request multiple candidates, or add per-run guidance.
  • Git hook integration — generate messages automatically from prepare-commit-msg while preserving and chaining an existing hook.
  • Live model discovery — search provider model catalogs with cgen model or from the configuration editor.
  • Safety guards — sensitive files and high-confidence secret patterns are blocked from the LLM payload; oversized diffs and large stagings require explicit confirmation.
  • Conventional Commits by default — optional gitmoji, custom templates, and commit messages in any language via ACR_LOCALE.
  • Presets & fallback — save provider configurations and fall back through them automatically when the primary LLM fails.
  • History tools — regenerate past commit messages (cgen alter), undo the latest commit safely (cgen undo), browse AI-generated commits (cgen history).
  • Self-updating — checksum-verified updates via cgen update or opt-in auto-update.

Why Rust?

Tools like opencommit do the same thing but require Node.js and weigh in at ~100MB of node_modules. cgen is a roughly 3MB self-contained executable. GNU/Linux release builds use the platform C library; a musl artifact is also published for portable x86_64 Linux installs.

cgen opencommit
Install size ~2 MB ~100 MB
Runtime deps None Node.js
Startup time Instant ~300ms (Node cold start)
Generation time ~800ms ~4s
Distribution Single binary npm install

Install

# Linux / macOS
curl -fsSL https://raw.githubusercontent.com/gtkacz/smart-commit-rs/main/scripts/install.sh | bash

# Windows (PowerShell)
irm https://raw.githubusercontent.com/gtkacz/smart-commit-rs/main/scripts/install.ps1 | iex

# Cargo
cargo install auto-commit-rs

→ Full installation instructions (manual download, custom install directory, binaries per platform).

Quick Start

# 1. Set your API key (one-time)
cgen config
# or: export ACR_API_KEY=your-key-here

# 2. Stage files and generate commit
git add .
cgen

→ Quick start walkthrough (what each step does, the review menu, push behavior).

Usage

cgen                    # Generate commit message and commit
cgen --dry-run          # Generate and show message without committing
cgen --all              # Stage tracked modifications/deletions, then generate
cgen --stdout           # Print only the generated message (for scripts)
cgen --generate 3       # Generate and choose from three candidates
cgen --prompt "focus on the API compatibility impact"
cgen --tag              # Also create the next semantic version tag
cgen config             # Interactive config editor (auto-detects scope)
cgen model              # Discover and select a provider model
cgen hook install       # Install repository-local Git hook integration
cgen alter <hash>       # Regenerate message from that commit's diff and rewrite it
cgen undo               # Undo latest commit with safety prompts (soft reset)
cgen history            # Browse AI-generated commits for the current repo

Any arguments passed to cgen (without a subcommand) are forwarded directly to git commit.

→ Full command reference (all flags, per-run overrides, diff filters, remaining subcommands).

Configuration

All settings use the ACR_ prefix and resolve in layers: defaults → global TOML → local .env in the repo root → process environment → CLI --set (highest priority, this run only). The settings you'll touch most:

Variable Default Description
ACR_PROVIDER groq LLM provider (see Providers)
ACR_MODEL llama-3.3-70b-versatile Model name
ACR_API_KEY unset API key (required by cloud providers)
ACR_LOCALE en Commit message language
ACR_ONE_LINER 1 Single-line commits (1/0)
ACR_USE_GITMOJI 0 Prepend gitmoji to messages (1/0)
ACR_REVIEW_COMMIT 1 Review message before committing (1/0)
ACR_POST_COMMIT_PUSH ask Push after commit (never/ask/always)

→ Full settings reference covers every variable, plus per-invocation overrides, variable interpolation, diff exclusion patterns, safety & workflow controls, and updating.

Providers

Built-in providers: Groq (default), OpenAI, Anthropic, Gemini, Grok, DeepSeek, OpenRouter, Mistral, Together, Fireworks, Perplexity, LM Studio, Ollama — plus any OpenAI-compatible custom endpoint.

→ Provider list and default models, presets, and fallback order.

Documentation

The full documentation lives at gtkacz.github.io/smart-commit-rs — installation, the complete configuration and command reference, provider details, and internals such as prompt design and commit history tracking.

AI-Generated Code Disclaimer

The majority of the code in this repository was generated by agentic AI. Every pull request and architecture decision is reviewed and refined by a human developer, and the codebase is gated by comprehensive unit tests and CI (coverage, formatting, lints, dependency audit, tests on Linux/macOS/Windows). The software is nevertheless provided "as is", without warranty of any kind.

Contributing

Contributions are welcome! Whether it's a new provider (often just 5 lines), a bug fix, or a documentation improvement, every bit helps.

See CONTRIBUTING.md for the development setup, the quality gates CI enforces, and a step-by-step guide to adding a new default provider.

License

MIT