GitWhisper VS Code Extension

🤖 AI-powered Git commit message generator for VS Code

Based on the original GitWhisper CLI tool by iamngoni
📖 Documentation | 🛠️ CLI Repository

GitWhisper is a VS Code extension that generates meaningful, conventional commit messages using AI. It supports multiple AI providers and languages, making your Git workflow more efficient and consistent.

✨ Features

🤖 AI-Powered Intelligence

• Multiple AI Models: Support for OpenAI, Claude, Gemini, Grok, Llama, DeepSeek, GitHub Models, and Ollama
• Smart Commit Messages: Generate conventional commit messages that follow best practices
• Change Analysis: AI-powered analysis of your code changes with detailed insights

⚡ Streamlined Workflow

• One-Click Actions: Generate, commit, and push to GitHub in seconds
• Smart Git Integration: Automatic staging, multi-repository support, and intelligent change detection
• Status Bar Integration: Quick access and real-time status updates

🌍 Flexible & Secure

• Multi-Language Support: Generate commit messages in 30+ languages
• Secure Storage: Uses VS Code's secure secrets storage for API keys and GitHub PATs
• Local AI Option: Use Ollama for completely offline, private AI processing
• Customizable Settings: Flexible configuration for models, languages, and workflow preferences

🔧 Developer-Friendly

• GitHub Integration: Seamless commit and push with Personal Access Token support
• File Filtering: Smart exclusion of lock files, logs, and temporary files
• Multi-Repository: Works across multiple Git repositories in your workspace
• Extension Integration: Works alongside your existing VS Code workflow

🚀 Quick Start

1. Install the Extension

   • Search for "GitWhisper" in VS Code Extensions (Ctrl+Shift+X)
   • Or install from the VS Code Marketplace (when published)

2. Set Up Your AI Provider

   • Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P)
   • Run GitWhisper: Set API Key
   • Choose your preferred AI provider and enter your API key
   • For Ollama (local AI), no API key needed - just install Ollama locally

3. Generate Your First Commit Message

   • Stage your changes: git add . or use VS Code's Source Control panel
   • Open Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run GitWhisper: Generate Commit Message
   • Review and commit!

4. Optional: Configure for Your Workflow

   • Run GitWhisper: Configure to set your preferred language, model, and workflow options
   • For GitHub repos, optionally set up a Personal Access Token for one-click push

📱 Visual Guide

Extension Interface Overview

The image above shows key GitWhisper interface elements in VS Code: status bar integration, source control panel icons, and command access points.

Where to Find GitWhisper

• Status Bar: Look for the GitWhisper indicator in the bottom status bar
• Source Control Panel: Find GitWhisper icons in the SCM toolbar
• Command Palette: Search for "GitWhisper" commands (Ctrl+Shift+P / Cmd+Shift+P)

🎯 Commands

All commands are accessible through the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) - search for "GitWhisper":

Command	Description
GitWhisper: Generate Commit Message	Generate AI-powered commit message
GitWhisper: Analyze Changes	Analyze staged changes with AI
GitWhisper: Set API Key	Configure API keys for AI providers
GitWhisper: Set GitHub Personal Access Token	Configure GitHub PAT for private repos
GitWhisper: Select Model	Choose default AI model and variant
GitWhisper: Set Language	Set commit message language
GitWhisper: Set Custom Ollama Model	Configure custom Ollama model variants
GitWhisper: Configure	Open configuration options
GitWhisper: Open Settings	Open extension settings panel

⚙️ Configuration

AI Providers

GitWhisper supports multiple AI providers:

• OpenAI (GPT-4o, GPT-4o Mini, GPT-4 Turbo, GPT-4, GPT-3.5 Turbo)
• Anthropic Claude (Claude 3.5 Sonnet, Claude 3.5 Haiku, Claude 3 Opus, Claude 3 Sonnet, Claude 3 Haiku)
• Google Gemini (Gemini 1.5 Pro, Gemini 1.5 Flash, Gemini Pro, Gemini Pro Vision)
• xAI Grok (Grok Beta, Grok Vision Beta)
• Meta Llama (Llama 3.2 90B, Llama 3.2 11B, Llama 3.1 70B, Llama 3.1 8B)
• DeepSeek (DeepSeek Chat, DeepSeek Coder)
• GitHub Models (GPT-4o, GPT-4o Mini, o1 Preview, o1 Mini)
• Ollama (Llama 3.2, Llama 3.1, Code Llama, Mistral, Mixtral, Qwen 2.5, DeepSeek Coder, Phi-3, Gemma 2, and custom models)

Settings

Configure GitWhisper through VS Code settings (Ctrl+, then search for "gitwhisper"):

Setting	Default	Description
gitwhisper.defaultModel	"openai"	Default AI model to use
gitwhisper.defaultVariant	"gpt-4o"	Default model variant
gitwhisper.language	"en;US"	Language for commit messages
gitwhisper.alwaysAdd	false	Auto-stage unstaged changes
gitwhisper.autoPush	false	Auto-push commits after creation
gitwhisper.ollamaBaseUrl	"http://localhost:11434"	Ollama server URL
gitwhisper.customOllamaVariant	""	Custom Ollama model name
gitwhisper.ignoredFiles	[...]	File patterns to ignore

Example configuration:

{
  "gitwhisper.defaultModel": "openai",
  "gitwhisper.defaultVariant": "gpt-4o",
  "gitwhisper.language": "en;US",
  "gitwhisper.alwaysAdd": false,
  "gitwhisper.autoPush": false,
  "gitwhisper.ollamaBaseUrl": "http://localhost:11434",
  "gitwhisper.customOllamaVariant": ""
}

Language Support

Generate commit messages in multiple languages (30+ supported):

Format: Use language code format "code;countryCode" (e.g., "en;US", "es;ES", "fr;FR")

• European: English, Spanish, French, German, Italian, Portuguese, Dutch, Swedish, Norwegian, Danish, Finnish, Polish, Czech, Hungarian, Romanian, Bulgarian, Greek, Turkish, Croatian, Serbian, Slovak, Slovenian, Lithuanian, Latvian, Estonian, Ukrainian, Russian
• Asian: Chinese (Simplified), Japanese, Korean, Thai, Vietnamese, Indonesian, Malay
• Middle Eastern: Arabic, Hebrew
• Indian Subcontinent: Hindi

🔧 API Key Setup

OpenAI

1. Visit OpenAI API Keys
2. Create a new API key
3. Set it in GitWhisper: GitWhisper: Set API Key → openai

Anthropic Claude

1. Visit Anthropic Console
2. Generate an API key
3. Set it in GitWhisper: GitWhisper: Set API Key → claude

Google Gemini

1. Visit Google AI Studio
2. Create an API key
3. Set it in GitWhisper: GitWhisper: Set API Key → gemini

xAI Grok

1. Visit xAI Console
2. Generate an API key
3. Set it in GitWhisper: GitWhisper: Set API Key → grok

Meta Llama

1. Visit Meta Llama
2. Get API access through approved providers
3. Set it in GitWhisper: GitWhisper: Set API Key → llama

DeepSeek

1. Visit DeepSeek Platform
2. Create an API key
3. Set it in GitWhisper: GitWhisper: Set API Key → deepseek

GitHub Models

1. Visit GitHub Models
2. Generate a Personal Access Token with model access
3. Set it in GitWhisper: GitWhisper: Set API Key → github

Ollama (Local)

1. Install Ollama
2. Pull a model: ollama pull llama3.2
3. No API key needed - uses local endpoint
4. Configure custom models in GitWhisper settings if needed

Troubleshooting Ollama

If Ollama commit generation is not working:

1. Check Ollama is running: Ensure Ollama is installed and running

   ollama serve
   

2. Verify model is available: Make sure the model is pulled

   ollama list
   ollama pull llama3.2  # or your desired model
   

3. Test Ollama API: Verify the API is accessible

   curl http://localhost:11434/api/generate -d '{
     "model": "llama3.2",
     "prompt": "Hello",
     "stream": false
   }'
   

4. Check VS Code Developer Console: Open Developer Tools (Help > Toggle Developer Tools) and check console for detailed error messages

5. Custom Base URL: If Ollama runs on different port/host, update base URL in GitWhisper settings

6. Verbose Output: If your model (like DeepSeek R1) generates too much text including thinking process, GitWhisper automatically cleans the response to extract just the commit message

   Setting Custom Ollama Models: Use GitWhisper: Set Custom Ollama Model command to configure specific models like deepseek-r1:1.5b, qwen2.5:7b, etc.

🔗 GitHub Integration

GitWhisper provides seamless GitHub integration with one-click commit and push functionality.

Setting up GitHub Personal Access Token

For private repositories or enhanced security, you can configure a GitHub Personal Access Token:

1. Generate a PAT:

   • Go to GitHub Settings > Developer settings > Personal access tokens
   • Click "Generate new token (classic)"
   • Select scopes: repo (for private repos) or public_repo (for public repos)
   • Copy the generated token

2. Configure in GitWhisper:

   • Run GitWhisper: Set GitHub Personal Access Token
   • Paste your token (it will be stored securely)

Commit and Push Workflow

When GitWhisper detects a GitHub repository, you'll see an additional "Commit and Push" button:

1. Generate commit message as usual
2. Choose your action:
   • Commit: Standard commit (local only)
   • Commit and Push: Commit and immediately push to GitHub
   • Copy to Clipboard: Copy message for manual use

The extension automatically handles:

• GitHub repository detection
• HTTPS/SSH URL conversion
• Secure PAT authentication
• Error handling and fallbacks

📝 Usage Examples

Basic Workflow

# Make your changes
git add .

Then open Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and run GitWhisper: Generate Commit Message

Advanced Workflows

Analyze Before Committing:

1. Stage your changes
2. Run GitWhisper: Analyze Changes to analyze changes
3. Run GitWhisper: Generate Commit Message to generate commit message
4. Review and commit

One-Click Commit and Push (GitHub repos):

1. Generate commit message as usual
2. Choose "Commit and Push" when prompted
3. GitWhisper handles authentication and pushing automatically

With Prefix Support

GitWhisper supports commit prefixes for ticket tracking:

JIRA-123: feat: add user authentication system
TICKET-456: fix: resolve memory leak in data processing

Multi-Repository Support

GitWhisper automatically detects and handles multiple Git repositories in your workspace.

🔧 Troubleshooting

Common Issues

"No staged changes found"

• Make sure you've staged your changes with git add . or through VS Code's Source Control panel
• Check that you're in a Git repository

API Key Issues

• Verify your API key is correct using GitWhisper: Set API Key
• Check your API provider's console for usage limits or billing issues
• For GitHub Models, ensure your PAT has the correct permissions

Ollama Not Working

• Ensure Ollama is installed and running: ollama serve
• Check if your model is available: ollama list
• Pull the model if needed: ollama pull llama3.2

Extension Not Activating

• Reload VS Code window: Ctrl+Shift+P → "Developer: Reload Window"
• Check VS Code's Output panel for error messages
• Ensure you have the minimum VS Code version (1.74.0+)

Getting Help

• Check the Issues page for known problems
• Open Developer Tools (Help > Toggle Developer Tools) to see console errors
• Review VS Code's Output panel for GitWhisper logs

❓ Frequently Asked Questions

Q: Which AI provider should I choose? A: For best results, we recommend GPT-4o (OpenAI) or Claude 3.5 Sonnet. For local/private use, try Ollama with Llama 3.2.

Q: Can I use GitWhisper without an internet connection? A: Yes! Install Ollama locally and use it with GitWhisper for completely offline AI-powered commit messages.

Q: Does GitWhisper work with private repositories? A: Absolutely! GitWhisper works with any Git repository. For GitHub private repos, set up a Personal Access Token for enhanced features.

Q: Can I customize the commit message format? A: GitWhisper follows conventional commit standards by default. The AI adapts to your project's existing commit style automatically.

Q: Is my code sent to AI providers? A: GitWhisper only sends git diff information (what changed), not your entire codebase. For maximum privacy, use Ollama (local AI).

Q: How do I change the default AI model? A: Use GitWhisper: Select Model command to open configuration.

Q: Does GitWhisper support multiple languages for commit messages? A: Yes! GitWhisper supports 30+ languages. Use GitWhisper: Set Language to configure your preferred language.

🛠️ Development

Prerequisites

• Node.js 18+
• pnpm
• VS Code 1.102.0 or higher

Setup

git clone https://github.com/emacliam/gitwhisper-vscode
cd gitwhisper-vscode
pnpm install

Build

pnpm run compile

Watch Mode

pnpm run watch

Package

pnpm run package

Testing

pnpm run test

Development Mode

Press F5 in VS Code to launch a new Extension Development Host window with your extension loaded.

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

📄 License

MIT License - see LICENSE file for details.

🙏 Acknowledgments

• Based on the original GitWhisper CLI tool by iamngoni
• Inspired by conventional commit standards
• Built with VS Code Extension API
• Uses multiple AI providers for enhanced flexibility

📞 Support

• 🐛 Report Issues
• 💬 Discussions
• 📧 Email Support

---

Happy Committing! 🚀