A git prepare-commit-msg hook for authoring commit messages with GPT-3. With this tool, you can easily generate clear, comprehensive and descriptive commit messages letting you focus on writing code.

See announcement blog post.

1. Install this tool locally with cargo (recommended).

cargo install --locked gptcommit

or on macOS, use homebrew

brew install zurawiki/brews/gptcommit

2. In your git repository, run the following command to install gptcommit as a git prepare-commit-msg hook. You will need to provide an OpenAI API key to complete the installation.

gptcommit install

To use gptcommit, simply run git commit as you normally would. The hook will automatically generate a commit message for you using a large language model like GPT. If you're not satisfied with the generated message, you can always edit it before committing.

Note: By default, gptcommit uses the GPT-3 model. Please ensure you have sufficient credits in your OpenAI account to use it.

gptcommit supports a number of configuration options that are read from $HOME/.config/gptcommit/config.toml. Configs are applied in the following order:

• User settings as read from $HOME/.config/gptcommit/config.toml.
• The settings as read from the repo clone at $GIT_ROOT/.git/gptcommit.toml.
• Environment variables starting with GPTCOMMIT__*.

See all the config options available with gptcommit config keys.

Persist your OpenAI key

gptcommit config set openai.api_key sk-...

or set it just for you local repo:

gptcommit config set --local openai.api_key sk-...

You can also config this setting via the GPTCOMMIT__OPENAI__API_KEY.

To maintain compatibility with other OpenAI clients, we support the OPENAI_API_KEY environment variables. This will take the highest precedence.

gptcommit uses text-davinci-003 by default. The model can be configured to use other models as below

gptcommit config set openai.model text-davinci-002

You can also config this setting via the GPTCOMMIT__OPENAI__MODEL.

For a list of public OpenAI models, checkout the OpenAI docs. You can also bring in your own fine-tuned model.

gptcommit uses English by default. The language can be configured to use other languages as below

gptcommit config set output.lang zh-cn

Now, supported languages are:

gptcommit config set allow-amend true

In the current design, gptcommit issues N+2 prompts, where N is the number of modified files with diffs under the max_token_limit. The other prompts are the title and summary.

OpenAI Completions are billed by "tokens" that are both sent and generated. Pricing per token depends on the model used. The number of tokens generated are generally predictable (as a commit message is usually only so big) but gptcommit could be sending over a lot of tokens in the form of diff data.

Today, I see two low-hanging solutions for reducing cost:

• Switch to a different model using the openai.model configuration option
• Reduce the side of prompts and diff data sent to OpenAI

OpenAI's pricing page can be found at https://openai.com/api/pricing/#faq-completions-pricing

By default, the githook is only run for new commits. If a template is set or the commit is being amended, the githook will skip by default.

Because the githook detected the user is supplying their own template, we make sure not to overwrite it with GPT. You can remove the commit template by making sure git config --local commit.template is blank.

You can allow gptcommit to summarize amended commits with the following configuration above.

You'll need to install Rust and the cargo toolchain first. Remember to configure your API key.

curl https://sh.rustup.rs -sSf | sh
bash
cargo install --locked gptcommit

# insert your openai api key https://platform.openai.com/account/api-keys
gptcommit config set openai.api_key # sk-...

All of these awesome projects are built using gptcommit.

• A VSCode extension you can install here | GitHub

If you encounter any bugs or have any suggestions for improvements, please open an issue on the repository.

This project is licensed under the MIT License.