Gemini Code Completion

AI-powered inline code completion for VS Code using Google's Gemini API.

Features

• 🚀 Real-time inline completions powered by Google Gemini
• 🧠 Smart context-aware suggestions with configurable context size
• ⚡ Fast with built-in completion caching
• 🔧 Configurable - choose your model, temperature, tokens, and more
• 🎯 Works with any language supported by VS Code
• 🔐 Secure - your API key is stored locally in VS Code settings

Prerequisites

1. Google Account with access to Google AI Studio
2. API Key - Get a free Gemini API key from aistudio.google.com
3. VS Code 1.82.0 or later

Installation

1. Clone or download this repository
2. Install dependencies:

   npm install
   

3. Build the extension:

   npm run compile
   

4. Package for local testing:

   vsce package
   

5. Install the .vsix file in VS Code: Extensions → Install from VSIX

Or install directly from VS Code Extensions Marketplace (when published).

Quick Start

1. Set API Key:

   • Run command: Gemini: Set API Key
   • Paste your Gemini API key (starts with AIza...)
   • Key is stored securely in VS Code settings

2. Start using completions:

   • Type code in any file
   • Completion suggestions appear after the debounce delay (default 500ms)
   • Press Tab or Enter to accept
   • Press Esc to dismiss

Configuration

Access settings via VS Code Settings (Cmd+, or Ctrl+,):

Core Settings

Generation Settings

Available Models

• gemini-2.5-flash - Fast, free tier (recommended for coding)
• gemini-2.5-pro - More capable, premium
• gemini-3-flash-preview - Latest Gemini 3, fast
• gemini-3.1-pro-preview - Most capable, premium only
• gemini-3.1-flash-lite-preview - Lightweight & efficient

Commands

Development

Building

npm run compile    # Compile TypeScript → JavaScript
npm run watch      # Watch for changes and compile
npm run lint       # Check code with ESLint

Debugging

1. Open the workspace in VS Code
2. Press F5 or go to Run → Start Debugging
3. A new VS Code window opens with the extension loaded
4. Set breakpoints and debug as normal
5. Output channel shows extension logs: Output → Gemini Code Completion

Testing

npm run test       # Run extension tests

Project Structure

.
├── src/
│   └── extension.ts         # Main extension code
├── out/                     # Compiled JavaScript (generated)
├── .vscode/
│   ├── launch.json         # Debug configuration
│   ├── tasks.json          # Build tasks
│   └── settings.json       # Workspace settings
├── package.json            # Project manifest
├── tsconfig.json           # TypeScript config
└── README.md               # This file

Troubleshooting

"No API key configured"

• Run Gemini: Set API Key command
• Verify the key starts with AIza...

Completions not appearing

1. Check the Output → Gemini Code Completion channel for errors
2. Verify your API key is valid
3. Check that geminiCodeCompletion.enabled is true
4. Try selecting a different model with Gemini: Select Model

Rate limit errors

• Wait a moment before triggering more completions
• Switch to a different model
• Check your API quota at aistudio.google.com

Slow completions

• Increase debounceDelay to reduce requests
• Reduce contextLines to send less context
• Try gemini-2.5-flash instead of gemini-2.5-pro

Security & Privacy

• Your API key is never sent anywhere except to Google's servers
• Keys are stored securely in VS Code's local settings
• No analytics or telemetry is collected by this extension
• All completions requests go directly to Google's Generative AI API

License

MIT

Support & Contributing

• Report issues: GitHub Issues
• Contributing: Pull requests welcome!

Related

• Google Generative AI SDK
• VS Code Extension API
• Google AI Studio