Git Commit AI — AI commit message generator CLI

![npm version](https://www.npmjs.com/package/@vavasilva/git-commit-ai)
![npm downloads](https://www.npmjs.com/package/@vavasilva/git-commit-ai)
![License: MIT](https://opensource.org/licenses/MIT)

> Generate Conventional Commit messages from staged git changes using LLMs (Ollama, llama.cpp, OpenAI, Anthropic, Groq).

git-commit-ai is a CLI that analyzes your git diff --staged and suggests high-quality Conventional Commits (type(scope): subject) with an interactive confirm/edit/regenerate flow.

Backends: Ollama (local), llama.cpp (local), OpenAI (GPT models), Anthropic (Claude), Groq (Llama)

!Demo

Quick Start

``bash

`Install`


npm install -g @vavasilva/git-commit-ai
1. Make changes to your code

echo "console.log('hello')" > hello.js
2. Stage your changes

git add hello.js
3. Generate commit message and commit

git-commit-ai
Output:

📝 Generated commit message

  feat: add hello.js script

[C]onfirm  [E]dit  [R]egenerate  [A]bort? c

✓ Committed: feat: add hello.js script


How it works

1. You stage your changes (git add ...) 2. git-commit-ai readsgit diff --staged3. A selected LLM backend proposes a Conventional Commit message 4. You confirm, edit, regenerate, or abort (no commit happens until you confirm)

`Features`

- Multiple Backends - Ollama (local), llama.cpp (local), OpenAI, Anthropic Claude, Groq - Auto-Detection - Automatically selects available backend - Conventional Commits - Generatestype(scope): subjectformat (Karma compatible) - Interactive Flow - Confirm, Edit, Regenerate, or Abort before committing - Individual Commits - Option to commit each file separately - Dry Run - Preview messages without committing - Git Hook - Auto-generate messages ongit commit- Summarize - Preview changes in plain English before committing - Debug Mode - Troubleshoot LLM responses - Configurable - Customize model, temperature, and more via config file

`Installation`

`bash

`Requires Node.js 20+`


npm install -g @vavasilva/git-commit-ai


$3
Choose at least one backend:

Ollama (Local, Free)`bash

`macOS`


brew install ollama
brew services start ollama
Linux

curl -fsSL https://ollama.com/install.sh | sh
sudo systemctl start ollama
Windows - download installer from:

https://ollama.com/download/windows
Pull a model (all platforms)

ollama pull llama3.1:8b


llama.cpp (Local, Free, Low Memory)

Run local GGUF models with llama-server (auto-detected on port 8080):

`bash

`Install llama.cpp`


macOS

brew install llama.cpp
Linux (Ubuntu/Debian) - build from source

sudo apt install build-essential cmake
git clone https://github.com/ggml-org/llama.cpp && cd llama.cpp
cmake -B build && cmake --build build --config Release
sudo cp build/bin/llama-server /usr/local/bin/
Windows - download pre-built binaries from:

https://github.com/ggml-org/llama.cpp/releases
Start the server (downloads model automatically from Hugging Face)

llama-server -hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080
Use with git-commit-ai (auto-detected if running on port 8080)

git-commit-ai
Or explicitly use llamacpp backend

git-commit-ai --backend llamacpp
Configure as default backend

git-commit-ai config --set backend=llamacpp


Run llama-server as a service


macOS (launchd)

`bash

`Create launchd service`


cat > ~/Library/LaunchAgents/com.llamacpp.server.plist << 'EOF'




    Label
    com.llamacpp.server
    ProgramArguments
    
        /opt/homebrew/bin/llama-server
        -hf
        Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF
        -ngl
        99
        --port
        8080
    
    RunAtLoad
    
    KeepAlive
    
    StandardOutPath
    /tmp/llama-server.log
    StandardErrorPath
    /tmp/llama-server.err


EOF
Start the service

launchctl load ~/Library/LaunchAgents/com.llamacpp.server.plist
Stop the service

launchctl unload ~/Library/LaunchAgents/com.llamacpp.server.plist
Check logs

tail -f /tmp/llama-server.log


Linux (systemd)

`bash

`Create systemd service`


sudo cat > /etc/systemd/system/llama-server.service << 'EOF'
[Unit]
Description=llama.cpp Server
After=network.target
[Service]
Type=simple
User=$USER
ExecStart=/usr/local/bin/llama-server -hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080
Restart=on-failure
RestartSec=10
StandardOutput=append:/var/log/llama-server.log
StandardError=append:/var/log/llama-server.err
[Install]
WantedBy=multi-user.target
EOF
Replace $USER with your username

sudo sed -i "s/\$USER/$USER/" /etc/systemd/system/llama-server.service
Enable and start the service

sudo systemctl daemon-reload
sudo systemctl enable llama-server
sudo systemctl start llama-server
Check status

sudo systemctl status llama-server
View logs

journalctl -u llama-server -f


Windows (Task Scheduler)
Option 1: PowerShell script with Task Scheduler

1. Create a startup script C:\llama-server\start-llama.ps1:`powershell

`start-llama.ps1`


Start-Process -FilePath "C:\llama-server\llama-server.exe"

-ArgumentList "-hf", "Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF", "-ngl", "99", "--port", "8080"


    -WindowStyle Hidden

-RedirectStandardOutput "C:\llama-server\llama-server.log"


    -RedirectStandardError "C:\llama-server\llama-server.err"

2. Create a scheduled task (run in PowerShell as Administrator):`powershell $action = New-ScheduledTaskAction -Execute "powershell.exe"
-Argument "-ExecutionPolicy Bypass -File C:\llama-server\start-llama.ps1"
$trigger = New-ScheduledTaskTrigger -AtStartup
$principal = New-ScheduledTaskPrincipal -UserId "$env:USERNAME" -LogonType S4U
Register-ScheduledTask -TaskName "LlamaServer" -Action $action -Trigger $trigger -Principal $principal

Start immediately

Start-ScheduledTask -TaskName "LlamaServer"

Stop the service

Stop-ScheduledTask -TaskName "LlamaServer"

Remove the service

Unregister-ScheduledTask -TaskName "LlamaServer" -Confirm:$false
``


Option 2: Using NSSM (Non-Sucking Service Manager)

`powershell

`Install NSSM (using chocolatey)`


choco install nssm
Install llama-server as a Windows service

nssm install LlamaServer "C:\llama-server\llama-server.exe" "-hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080"
nssm set LlamaServer AppDirectory "C:\llama-server"
nssm set LlamaServer AppStdout "C:\llama-server\llama-server.log"
nssm set LlamaServer AppStderr "C:\llama-server\llama-server.err"
Start the service

nssm start LlamaServer
Stop the service

nssm stop LlamaServer
Remove the service

nssm remove LlamaServer confirm

OpenAI`bash export OPENAI_API_KEY="your-api-key"`

OpenAI-Compatible APIs

Any OpenAI-compatible API can be used by setting OPENAI_BASE_URL:`bash

`Local server (llama.cpp, vLLM, etc.)`


export OPENAI_BASE_URL="http://localhost:8080/v1"
Or other providers (Together AI, Anyscale, etc.)

export OPENAI_BASE_URL="https://api.together.xyz/v1"
export OPENAI_API_KEY="your-api-key"

Anthropic (Claude)`bash export ANTHROPIC_API_KEY="your-api-key"`

Groq (Fast & Free tier)`bash export GROQ_API_KEY="your-api-key"`

`Usage`

`bash

`Basic: stage files + generate + confirm + commit`


git add file1.ts file2.ts
git-commit-ai
Stage all changes and commit

git-commit-ai --all
Auto-commit without confirmation

git add .
git-commit-ai -y
Commit and push in one command

git add .
git-commit-ai --push
Stage all changes and commit (equivalent to git add . && git-commit-ai)

git-commit-ai --all
Commit each modified file separately

git-commit-ai --individual
Preview message without committing (dry run)

git add .
git-commit-ai --dry-run
Amend the last commit with a new message

git-commit-ai --amend
Force a specific scope and type

git-commit-ai --scope auth --type fix
Generate message in a specific language

git-commit-ai --lang pt
Reference an issue

git-commit-ai --issue 123
Mark as breaking change

git-commit-ai --breaking
Add co-authors

git-commit-ai --co-author "Jane Doe "
Provide additional context

git-commit-ai --context "This fixes the login bug reported by QA"
Use a specific backend

git-commit-ai --backend llamacpp
git-commit-ai --backend openai
git-commit-ai --backend anthropic
git-commit-ai --backend groq
Override model

git-commit-ai --model gpt-4o
git-commit-ai --model claude-3-sonnet-20240229
Adjust creativity (temperature)

git-commit-ai --temperature 0.3
Preview changes before committing

git add .
git-commit-ai summarize
Enable debug output for troubleshooting

git-commit-ai --debug
Show current config

git-commit-ai config
Set a config value

git-commit-ai config --set backend=llamacpp
git-commit-ai config --set model=gpt-4o
git-commit-ai config --set temperature=0.5
Use short aliases

git-commit-ai config --set lang=pt        # → default_language
git-commit-ai config --set scope=api      # → default_scope
git-commit-ai config --set type=feat      # → default_type
git-commit-ai config --set temp=0.5       # → temperature
List valid config keys and aliases

git-commit-ai config --list-keys
Create/edit config file manually

git-commit-ai config --edit


Git Hook (Auto-generate on commit)
Install a git hook to automatically generate commit messages:

`bash

`Install the hook`


git-commit-ai hook --install
Now just use git commit normally!

git add .
git commit
Message is auto-generated and opens in your editor
Check hook status

git-commit-ai hook --status
Remove the hook

git-commit-ai hook --remove


Interactive Flow

`📝 Generated commit message

feat(auth): add login validation

[C]onfirm [E]dit [R]egenerate [A]bort? _`

`Configuration`

`$3`

Location: ~/.config/git-commit-ai/config.toml

`toml

`Backend: ollama, llamacpp, openai, anthropic, groq`


backend = "ollama"
model = "llama3.1:8b"
ollama_url = "http://localhost:11434"
temperature = 0.7
retry_temperatures = [0.5, 0.3, 0.2]
OpenAI Base URL - change this to use OpenAI-compatible APIs

Examples:

  - Default OpenAI: https://api.openai.com/v1

  - llama.cpp:      http://localhost:8080/v1

  - Together AI:    https://api.together.xyz/v1

openai_base_url = "https://api.openai.com/v1"
Optional: Ignore files from diff analysis

ignore_patterns = [".lock", "package-lock.json", ".min.js"]
Optional: Set defaults for commit messages

default_scope = "api"          # Default scope if not specified
default_type = "feat"          # Default commit type
default_language = "en"        # Default language (en, pt, es, fr, de)

$3

Create .gitcommitai or .gitcommitai.toml in your project root to override global settings:

`toml

`.gitcommitai`


default_scope = "frontend"
default_language = "pt"
ignore_patterns = ["dist/", ".generated.ts"]


$3
| Backend | Default Model |
|---------|---------------|
| ollama | llama3.1:8b |
| llamacpp | gpt-4o-mini (alias) |
| openai | gpt-4o-mini |
| anthropic | claude-3-haiku-20240307 |
| groq | llama-3.1-8b-instant |
CLI Options

| Option | Description | |--------|-------------| |-a, --all| Stage all changes before committing | |-p, --push| Push after commit | |-y, --yes| Skip confirmation | |-i, --individual| Commit files individually | |-d, --debug| Enable debug output | |--dry-run| Show message without committing | |--amend| Regenerate and amend the last commit | |-b, --backend | Backend to use | |-m, --model | Override model | |-t, --temperature | Override temperature (0.0-1.0) | |-s, --scope | Force a specific scope (e.g., auth, api) | |--type | Force commit type (feat, fix, docs, etc.) | |-c, --context | Provide additional context for generation | |-l, --lang | Language for message (en, pt, es, fr, de) | | --issue | Reference an issue (e.g., 123 or #123) | | --breaking | Mark as breaking change (adds ! to type) | | --co-author | Add co-author (can be repeated) |

`Config Commands`

| Command | Description | |---------|-------------| | config | Show current configuration | | config --edit | Create/edit config file manually | | config --set | Set a config value | | config --list-keys | List all valid config keys |

`Commit Types (Conventional Commits)`

| Type | Description | |------|-------------| | feat | New feature | | fix | Bug fix | | docs | Documentation | | style | Formatting (no code change) | | refactor | Code restructuring | | test | Adding tests | | build | Build system or dependencies | | chore | Maintenance tasks |

`Environment Variables`

| Variable | Description | |----------|-------------| | OPENAI_API_KEY | OpenAI API key | | OPENAI_BASE_URL | OpenAI-compatible API base URL (default: https://api.openai.com/v1) | | ANTHROPIC_API_KEY | Anthropic API key | | GROQ_API_KEY` | Groq API key |

`License`

MIT

`Git Commit AI — AI commit message generator CLI`

![npm version](https://www.npmjs.com/package/@vavasilva/git-commit-ai) ![npm downloads](https://www.npmjs.com/package/@vavasilva/git-commit-ai) ![License: MIT](https://opensource.org/licenses/MIT)

> Generate Conventional Commit messages from staged git changes using LLMs (Ollama, llama.cpp, OpenAI, Anthropic, Groq).

git-commit-ai is a CLI that analyzes your git diff --staged and suggests high-quality Conventional Commits (type(scope): subject) with an interactive confirm/edit/regenerate flow.

Backends: Ollama (local), llama.cpp (local), OpenAI (GPT models), Anthropic (Claude), Groq (Llama)

!Demo

`Quick Start`

``bash

`Install`


npm install -g @vavasilva/git-commit-ai
1. Make changes to your code

echo "console.log('hello')" > hello.js
2. Stage your changes

git add hello.js
3. Generate commit message and commit

git-commit-ai
Output:

📝 Generated commit message

  feat: add hello.js script

[C]onfirm  [E]dit  [R]egenerate  [A]bort? c

✓ Committed: feat: add hello.js script

`
How it works
1. You stage your changes (git add ...)
2. git-commit-ai reads git diff --staged
3. A selected LLM backend proposes a Conventional Commit message
4. You confirm, edit, regenerate, or abort (no commit happens until you confirm)
Features
- Multiple Backends - Ollama (local), llama.cpp (local), OpenAI, Anthropic Claude, Groq
- Auto-Detection - Automatically selects available backend
- Conventional Commits - Generates type(scope): subject format (Karma compatible)
- Interactive Flow - Confirm, Edit, Regenerate, or Abort before committing
- Individual Commits - Option to commit each file separately
- Dry Run - Preview messages without committing
- Git Hook - Auto-generate messages on git commit
- Summarize - Preview changes in plain English before committing
- Debug Mode - Troubleshoot LLM responses
- Configurable - Customize model, temperature, and more via config file
Installation
`bash
Requires Node.js 20+

npm install -g @vavasilva/git-commit-ai
`
$3
Choose at least one backend:Ollama (Local, Free)
`bash
macOS

brew install ollama
brew services start ollama
Linux

curl -fsSL https://ollama.com/install.sh | sh
sudo systemctl start ollama
Windows - download installer from:

https://ollama.com/download/windows
Pull a model (all platforms)

ollama pull llama3.1:8b
`
llama.cpp (Local, Free, Low Memory)Run local GGUF models with llama-server (auto-detected on port 8080):
`bash
Install llama.cpp

macOS

brew install llama.cpp
Linux (Ubuntu/Debian) - build from source

sudo apt install build-essential cmake
git clone https://github.com/ggml-org/llama.cpp && cd llama.cpp
cmake -B build && cmake --build build --config Release
sudo cp build/bin/llama-server /usr/local/bin/
Windows - download pre-built binaries from:

https://github.com/ggml-org/llama.cpp/releases
Start the server (downloads model automatically from Hugging Face)

llama-server -hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080
Use with git-commit-ai (auto-detected if running on port 8080)

git-commit-ai
Or explicitly use llamacpp backend

git-commit-ai --backend llamacpp
Configure as default backend

git-commit-ai config --set backend=llamacpp
`
Run llama-server as a service

macOS (launchd)
`bash
Create launchd service

cat > ~/Library/LaunchAgents/com.llamacpp.server.plist << 'EOF'




    Label
    com.llamacpp.server
    ProgramArguments
    
        /opt/homebrew/bin/llama-server
        -hf
        Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF
        -ngl
        99
        --port
        8080
    
    RunAtLoad
    
    KeepAlive
    
    StandardOutPath
    /tmp/llama-server.log
    StandardErrorPath
    /tmp/llama-server.err


EOF
Start the service

launchctl load ~/Library/LaunchAgents/com.llamacpp.server.plist
Stop the service

launchctl unload ~/Library/LaunchAgents/com.llamacpp.server.plist
Check logs

tail -f /tmp/llama-server.log
`

Linux (systemd)
`bash
Create systemd service

sudo cat > /etc/systemd/system/llama-server.service << 'EOF'
[Unit]
Description=llama.cpp Server
After=network.target
[Service]
Type=simple
User=$USER
ExecStart=/usr/local/bin/llama-server -hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080
Restart=on-failure
RestartSec=10
StandardOutput=append:/var/log/llama-server.log
StandardError=append:/var/log/llama-server.err
[Install]
WantedBy=multi-user.target
EOF
Replace $USER with your username

sudo sed -i "s/\$USER/$USER/" /etc/systemd/system/llama-server.service
Enable and start the service

sudo systemctl daemon-reload
sudo systemctl enable llama-server
sudo systemctl start llama-server
Check status

sudo systemctl status llama-server
View logs

journalctl -u llama-server -f
`

Windows (Task Scheduler)
Option 1: PowerShell script with Task Scheduler1. Create a startup script C:\llama-server\start-llama.ps1:
`powershell
start-llama.ps1

Start-Process -FilePath "C:\llama-server\llama-server.exe" 
    -ArgumentList "-hf", "Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF", "-ngl", "99", "--port", "8080" 
    -WindowStyle Hidden 
    -RedirectStandardOutput "C:\llama-server\llama-server.log" 
    -RedirectStandardError "C:\llama-server\llama-server.err"
`2. Create a scheduled task (run in PowerShell as Administrator):
`powershell
$action = New-ScheduledTaskAction -Execute "powershell.exe" 
    -Argument "-ExecutionPolicy Bypass -File C:\llama-server\start-llama.ps1"
$trigger = New-ScheduledTaskTrigger -AtStartup
$principal = New-ScheduledTaskPrincipal -UserId "$env:USERNAME" -LogonType S4U
Register-ScheduledTask -TaskName "LlamaServer" -Action $action -Trigger $trigger -Principal $principal
Start immediately

Start-ScheduledTask -TaskName "LlamaServer"
Stop the service

Stop-ScheduledTask -TaskName "LlamaServer"
Remove the service

Unregister-ScheduledTask -TaskName "LlamaServer" -Confirm:$false
``
Option 2: Using NSSM (Non-Sucking Service Manager)`powershell
Install NSSM (using chocolatey)

choco install nssm
Install llama-server as a Windows service

nssm install LlamaServer "C:\llama-server\llama-server.exe" "-hf Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF -ngl 99 --port 8080"
nssm set LlamaServer AppDirectory "C:\llama-server"
nssm set LlamaServer AppStdout "C:\llama-server\llama-server.log"
nssm set LlamaServer AppStderr "C:\llama-server\llama-server.err"
Start the service

nssm start LlamaServer
Stop the service

nssm stop LlamaServer
Remove the service

nssm remove LlamaServer confirm
`
OpenAI
`bash
export OPENAI_API_KEY="your-api-key"
`
OpenAI-Compatible APIs
Any OpenAI-compatible API can be used by setting OPENAI_BASE_URL:
`bash
Local server (llama.cpp, vLLM, etc.)

export OPENAI_BASE_URL="http://localhost:8080/v1"
Or other providers (Together AI, Anyscale, etc.)

export OPENAI_BASE_URL="https://api.together.xyz/v1"
export OPENAI_API_KEY="your-api-key"
`Anthropic (Claude)
`bash
export ANTHROPIC_API_KEY="your-api-key"
`
Groq (Fast & Free tier)
`bash
export GROQ_API_KEY="your-api-key"
`
Usage
`bash
Basic: stage files + generate + confirm + commit

git add file1.ts file2.ts
git-commit-ai
Stage all changes and commit

git-commit-ai --all
Auto-commit without confirmation

git add .
git-commit-ai -y
Commit and push in one command

git add .
git-commit-ai --push
Stage all changes and commit (equivalent to git add . && git-commit-ai)

git-commit-ai --all
Commit each modified file separately

git-commit-ai --individual
Preview message without committing (dry run)

git add .
git-commit-ai --dry-run
Amend the last commit with a new message

git-commit-ai --amend
Force a specific scope and type

git-commit-ai --scope auth --type fix
Generate message in a specific language

git-commit-ai --lang pt
Reference an issue

git-commit-ai --issue 123
Mark as breaking change

git-commit-ai --breaking
Add co-authors

git-commit-ai --co-author "Jane Doe "
Provide additional context

git-commit-ai --context "This fixes the login bug reported by QA"
Use a specific backend

git-commit-ai --backend llamacpp
git-commit-ai --backend openai
git-commit-ai --backend anthropic
git-commit-ai --backend groq
Override model

git-commit-ai --model gpt-4o
git-commit-ai --model claude-3-sonnet-20240229
Adjust creativity (temperature)

git-commit-ai --temperature 0.3
Preview changes before committing

git add .
git-commit-ai summarize
Enable debug output for troubleshooting

git-commit-ai --debug
Show current config

git-commit-ai config
Set a config value

git-commit-ai config --set backend=llamacpp
git-commit-ai config --set model=gpt-4o
git-commit-ai config --set temperature=0.5
Use short aliases

git-commit-ai config --set lang=pt        # → default_language
git-commit-ai config --set scope=api      # → default_scope
git-commit-ai config --set type=feat      # → default_type
git-commit-ai config --set temp=0.5       # → temperature
List valid config keys and aliases

git-commit-ai config --list-keys
Create/edit config file manually

git-commit-ai config --edit
`
Git Hook (Auto-generate on commit)
Install a git hook to automatically generate commit messages:`bash
Install the hook

git-commit-ai hook --install
Now just use git commit normally!

git add .
git commit
Message is auto-generated and opens in your editor
Check hook status

git-commit-ai hook --status
Remove the hook

git-commit-ai hook --remove
`
Interactive Flow
`
📝 Generated commit message
  feat(auth): add login validation
[C]onfirm  [E]dit  [R]egenerate  [A]bort? _
`
Configuration
$3
Location: ~/.config/git-commit-ai/config.toml
`toml
Backend: ollama, llamacpp, openai, anthropic, groq

backend = "ollama"
model = "llama3.1:8b"
ollama_url = "http://localhost:11434"
temperature = 0.7
retry_temperatures = [0.5, 0.3, 0.2]
OpenAI Base URL - change this to use OpenAI-compatible APIs

Examples:

  - Default OpenAI: https://api.openai.com/v1

  - llama.cpp:      http://localhost:8080/v1

  - Together AI:    https://api.together.xyz/v1

openai_base_url = "https://api.openai.com/v1"
Optional: Ignore files from diff analysis

ignore_patterns = [".lock", "package-lock.json", ".min.js"]
Optional: Set defaults for commit messages

default_scope = "api"          # Default scope if not specified
default_type = "feat"          # Default commit type
default_language = "en"        # Default language (en, pt, es, fr, de)
`
$3
Create .gitcommitai or .gitcommitai.toml in your project root to override global settings:
`toml
.gitcommitai

default_scope = "frontend"
default_language = "pt"
ignore_patterns = ["dist/", ".generated.ts"]
`
$3
| Backend | Default Model |
|---------|---------------|
| ollama | llama3.1:8b |
| llamacpp | gpt-4o-mini (alias) |
| openai | gpt-4o-mini |
| anthropic | claude-3-haiku-20240307 |
| groq | llama-3.1-8b-instant |
CLI Options
| Option | Description |
|--------|-------------|
| -a, --all | Stage all changes before committing |
| -p, --push | Push after commit |
| -y, --yes | Skip confirmation |
| -i, --individual | Commit files individually |
| -d, --debug | Enable debug output |
| --dry-run | Show message without committing |
| --amend | Regenerate and amend the last commit |
| -b, --backend  | Backend to use |
| -m, --model  | Override model |
| -t, --temperature  | Override temperature (0.0-1.0) |
| -s, --scope  | Force a specific scope (e.g., auth, api) |
| --type  | Force commit type (feat, fix, docs, etc.) |
| -c, --context  | Provide additional context for generation |
| -l, --lang  | Language for message (en, pt, es, fr, de) |
| --issue  | Reference an issue (e.g., 123 or #123) |
| --breaking | Mark as breaking change (adds ! to type) |
| --co-author  | Add co-author (can be repeated) |
Config Commands
| Command | Description |
|---------|-------------|
| config | Show current configuration |
| config --edit | Create/edit config file manually |
| config --set  | Set a config value |
| config --list-keys | List all valid config keys |
Commit Types (Conventional Commits)
| Type | Description |
|------|-------------|
| feat | New feature |
| fix | Bug fix |
| docs | Documentation |
| style | Formatting (no code change) |
| refactor | Code restructuring |
| test | Adding tests |
| build | Build system or dependencies |
| chore | Maintenance tasks |
Environment Variables
| Variable | Description |
|----------|-------------|
| OPENAI_API_KEY | OpenAI API key |
| OPENAI_BASE_URL | OpenAI-compatible API base URL (default: https://api.openai.com/v1) |
| ANTHROPIC_API_KEY | Anthropic API key |
| GROQ_API_KEY` | Groq API key |
License
MIT