git-agent

git-agent writes git commit messages for your changes using a local LLM served through an OpenAI-compatible API (LM Studio, Ollama, llama.cpp, vLLM, or a hosted provider). You review each proposed message, then commit it, edit it, regenerate it, or tell the model what to change.

Why I made git-agent

I made git-agent mainly to get the commit message generation from IntelliJ IDEA without running IDEA. IDEA uses nearly 3-4 GB of RAM, and I need that memory for running local language models. git-agent also has stronger automation features, for use in pipelines or with lightweight code editors. The code was developed using Gemma 4 locally, and refactored with Claude.

Install

git-agent is published on crates.io. With Rust 1.88 or newer installed, run:

cargo install git-agent --locked

Cargo builds an optimized binary and copies it to ~/.cargo/bin/ (or $CARGO_HOME/bin/). --locked uses the exact dependency versions the release was tested with. Run the same command again to update to the latest release, and cargo uninstall git-agent to remove it.

Because the binary is called git-agent, git also runs it as git agent.

To install from a local checkout instead, for example to try unreleased changes:

cargo install --path . --locked

If you installed Rust with Homebrew or a Linux package manager

~/.cargo/bin must be on your PATH. rustup adds it for you, but Homebrew and Linux package managers don't, so add it yourself, for example in ~/.zshrc:

export PATH="$HOME/.cargo/bin:$PATH"

Usage

git agent                  # describe staged changes, then review
git agent --all            # stage everything (including untracked files) first
git agent -H "fixes #42"   # give the model extra context
git agent --commit         # commit right away, no review
git agent --push           # review, then push after committing
git agent --commit --push  # commit and push, no review
git agent --raw            # print just the message, e.g. for `| git commit -F -`
git agent -- --signoff     # pass extra arguments to `git commit`

If you have staged changes, git-agent describes and commits only those. If nothing is staged, it describes all changes to tracked files and commits them with git commit --all. Untracked files are left out unless you use --all.

After git-agent generates a message, you choose what to do with it:

If git commit fails (for example, because a hook rejects it), git-agent saves the message to .git/GIT_AGENT_MSG so you can reuse it with git commit --file .git/GIT_AGENT_MSG.

Pushing runs a plain git push, so it works with GitHub or any other remote and uses your existing git credentials. The first time you push a branch that has no upstream, git-agent sets one: it uses remote.pushDefault if you've configured it, otherwise your only remote, otherwise origin. If the repository has no remote yet, create one on GitHub with gh repo create --source . --push. If the push fails, the commit is kept; fix the problem and run git push.

Before pushing, git-agent fetches from the remote. If the remote has commits you don't have, such as a file you edited on GitHub, git-agent lists them and rebases your commit on top of them. That keeps the history linear and leaves the remote commits in place. The review menu asks before rebasing; --commit --push rebases without asking. Uncommitted changes are set aside during the rebase and restored afterwards.

If your commit conflicts with the remote commits, git-agent cancels the rebase and leaves your branch as it was. Run git pull --rebase, resolve the conflicts, and git push. git-agent never force-pushes.

Choosing a model

You don't have to configure a model. git-agent asks the server which model is loaded and uses it. If several models are loaded, or none is and the server has more than one, git-agent asks you to choose with --model or model in the config file. git agent models lists the server's models and marks the one git-agent uses.

git-agent also asks the server how large the model's context window is, so it knows how much of your changes it can send. LM Studio, Ollama, llama.cpp and vLLM report it. For other servers, and for models that aren't loaded yet, git-agent assumes 4,096 tokens and says so. Set context_length to tell it the real size.

Large changes

Before each run, git-agent prints the model it uses and how much of the context the prompt takes:

🤖 google/gemma-4-e4b (loaded on the server) · prompt about 24,578 of 131,072 context tokens

The prompt always leaves room for the reply. When your changes don't fit, git-agent shortens them in steps and stops as soon as they fit:

1. It leaves out the unchanged lines that git shows around each change.
2. It shares the remaining room between the files. Files with small diffs are sent whole, and the larger ones are cut short, so every file shows at least its first changes.
3. If there are too many files even for that, the largest ones are sent by name only. The list of changed files always includes them.

git-agent then tells you what it left out:

warning: the changes need about 24,656 tokens, more than fit in the context, so git-agent sent part of them:
   46 of 58 files in full, without the unchanged lines around the changes
   12 cut short: src/builder/arg.rs (6 of 23 changed lines), src/builder/command.rs (6 of 93 changed lines), and 10 more

When that happens, the message may miss details from the shortened files. You get a better message by committing in smaller parts: stage some of the files, run git agent, and repeat. Loading the model with a larger context also helps.

Configuration

Settings come from these sources, and later ones override earlier ones:

1. Built-in defaults
2. The config file: ~/.config/git-agent/config.toml (or $XDG_CONFIG_HOME/git-agent/config.toml)
3. Environment variables
4. Command-line flags

git agent config init       # write a commented config file with the defaults
git agent config show       # print the effective settings
git agent models            # list models on the server; * marks the configured one

To switch models for a single run:

git agent -m qwen/qwen3-8b
GIT_AGENT_BASE_URL=http://localhost:11434/v1 git agent -m llama3.2   # Ollama

Notes:

• Files matching exclude (by default Cargo.lock, package-lock.json, and similar) still appear in the list of changed files sent to the model, but their diffs don't. Deleted files are sent by name, without their old content.
• Large prompts take longer to process. To keep them smaller, set context_length below the model's real context.
• git-agent removes reasoning blocks (<think>…</think>) and code fences around the model's reply.
• Set NO_COLOR to turn off colored output.

Contributing

Contributions are welcome! Open an issue for bugs and feature ideas, or send a pull request. For larger changes, please open an issue first so we can discuss the approach.

Before you send a pull request, make sure these pass:

cargo fmt --check
cargo clippy --all-targets
cargo test

The tests create temporary git repositories and don't need a running model server.

Releasing

Maintainers release a new version with scripts/release.sh <version>. The script runs the checks, bumps the version in Cargo.toml, publishes the crate to crates.io, tags and pushes the release commit, and creates a GitHub release. Run it with --dry-run first to see each step without changing anything. It needs the gh CLI and a crates.io token from cargo login.

License

git-agent is licensed under the MIT License. By contributing, you agree that your contributions are licensed under the same terms.