🤖 Gemini Copilot

A lightweight, high-performance Visual Studio Code extension that provides AI-powered inline code completions (ghost text) using Google Gemini 3.6 Flash.

🚀 Features

• 🔒 Secure API Key Management (BYOK): Prompts for and securely stores your own Gemini API key using VS Code's SecretStorage API. The key is persisted in your operating system's native keychain (such as Windows Credential Manager or macOS Keychain), not in plain text files.
• ⚡ Configurable Keystroke Debouncing: Restricts API calls using a configurable delay (default 750ms) to avoid calling the Gemini API on every single keystroke.
• 🛑 In-Flight Request Cancellation: Automatically detects when you resume typing while a network request is still active and aborts the Gemini API call immediately using an AbortController, saving network bandwidth and rate limit quota.
• 🎯 Smart Code Continuation: Uses system instructions to guide gemini-3.6-flash to output ONLY raw code continuation. It automatically post-processes the model response to strip unwanted markdown formatting, notes, or conversational text.
• ⏳ Visual Rate Limit Cooldown: Captures HTTP 429 (Rate Limit Exceeded) errors, parses the suggested retry delay, and displays a real-time countdown timer in your status bar: $(alert) Gemini Copilot (Rate Limited - 45s).

🛠️ Tech Stack

• TypeScript - Primary development language.
• VS Code Extension API - Integrates directly with editor inline suggestion APIs.
• Node.js https Module - High-speed API queries using native network modules for maximum compatibility across different VS Code client versions.
• esbuild - Lightning-fast bundling and minification.

⚙️ Getting Started

Prerequisites

• VS Code: Version 1.75.0 or newer.
• Node.js: Version 18.x or newer (we recommend using Node.js 20.x or newer).
• Gemini API Key: Retrieve your own key from Google AI Studio.

Installation

1. Clone the Repository:

   git clone https://github.com/your-username/gemini-copilot.git
   cd gemini-copilot
   

2. Install Dependencies:

   npm install
   

3. Run the Extension Locally:

   • Open the cloned directory in VS Code.
   • Press F5 (or go to Run and Debug > Run Extension). This compiles the TypeScript code in the background and opens a new Extension Development Host window.

Environment Setup (Secure BYOK)

Instead of environment variables or settings.json, your API key is managed securely inside VS Code:

1. Inside the new Extension Development Host window, open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P).
2. Search and run the command: Gemini Copilot: Set Gemini API Key.
3. Paste your Gemini API key and press Enter.
4. The status bar in the bottom-right corner will update to show ✓ Gemini Copilot.

💡 Usage Examples

Inline Completion (Ghost Text)

1. Open any source code file (e.g., Python, JavaScript, TypeScript, Go).
2. Start typing a function definition or a logic block. For example:

   def find_prime_numbers(limit):
   

3. Pause typing for 750ms (or your configured debounce time).
4. The Gemini API returns the completion suggestion in grey ghost text:

   def find_prime_numbers(limit):
       # Ghost text begins here:
       primes = []
       for num in range(2, limit + 1):
           if all(num % i != 0 for i in range(2, int(num ** 0.5) + 1)):
               primes.append(num)
       return primes
   

5. Press Tab to accept the suggestion, or keep typing to cancel.

Custom Configuration

Adjust the extension behaviour by adding the following settings to your VS Code settings.json (or via the Settings UI):

{
  "geminiCopilot.debounceDelay": 1000
}

• geminiCopilot.debounceDelay (default 750): The amount of time in milliseconds to wait after the last keystroke before sending the request. Increase this if you frequently hit free-tier rate limits.

🧪 Running Tests & Inspection

Since inline completions are asynchronous, you can trace the extension behavior and logs in real-time:

1. Open the Output panel in the bottom of your VS Code workspace (Ctrl+Shift+U / Cmd+Shift+U).
2. In the dropdown menu at the top-right of the Output panel, select Gemini Copilot.
3. Here you will see live trace statements showing:
   • Live debounce timer start/stop events.
   • Active request aborts due to keystroke cancellation.
   • API request statuses, execution latencies, and suggestions.
   • Cool-down countdown progression when rate-limited.

🤝 Contributing

Contributions are welcome! Please open an issue or submit a pull request if you have ideas for improvements, bug fixes, or new features.

📄 License

This project is licensed under the MIT License.