gitmsg

gitmsg suggests a concise Conventional Commit message from your staged Git diff using an OpenAI-compatible API. It can talk to OpenAI or Google AI Studio with the same chat completions flow.

What it does

• Reads the staged diff with git diff --cached
• Sends the diff to an OpenAI-compatible chat completions endpoint
• Prints a suggested commit message in type(scope): description form
• Optionally runs git commit -m "..."

Installation from source

git clone https://github.com/jsicard833/gitmsg.git
cd gitmsg
cargo build --release

The binary is built as target/release/gitmsg. You can also install it locally with:

cargo install --path .

Configuration

Set these environment variables before running the tool:

• AI_PROVIDER - optional provider selector, openai or google; defaults to auto-detect
• OPENAI_API_KEY - required when using OpenAI
• OPENAI_BASE_URL - optional custom base URL, defaults to https://api.openai.com/v1
• OPENAI_MODEL - optional model name, defaults to gpt-5.4-mini
• GOOGLE_API_KEY - required when using Google AI Studio
• GOOGLE_BASE_URL - optional custom base URL, defaults to https://generativelanguage.googleapis.com/v1beta/openai
• GOOGLE_MODEL - optional model name, defaults to gemini-3.5-flash

Example:

export OPENAI_API_KEY=your_key_here
export OPENAI_BASE_URL=https://api.openai.com/v1
export OPENAI_MODEL=gpt-5.4-mini

# or

export AI_PROVIDER=google
export GOOGLE_API_KEY=your_google_key_here
export GOOGLE_MODEL=gemini-3.5-flash

Usage

Generate a suggested commit message:

git add .
gitmsg

Generate and commit immediately:

gitmsg --commit

Print debug details while running:

gitmsg --verbose

Interactive mode with three suggestions and a selection prompt:

gitmsg -i

After you choose a suggestion, gitmsg will commit it immediately.

You’ll see output like this:

1. feat(vector): add iterator support
2. feat(vector): implement begin/end methods
3. refactor(vector): improve iterator handling

Select [1]:

Press Enter to accept the first suggestion.

Example output

feat(parser): add staged diff summarization