Skip to content
| Marketplace
Sign in
Visual Studio Code>Programming Languages>AIAutopilotEngineNew to Visual Studio Code? Get it now.
AIAutopilotEngine

AIAutopilotEngine

Hamza Cantürk

| (0) | Free
Multi-pass self-refining LLM optimization engine for VS Code
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
Copied to clipboard
More Info

AIAutopilotEngine

🚀 Multi-pass Self-Refining LLM Optimization Engine for VS Code

AIAutopilotEngine is a VS Code Chat Participant that automatically improves your code and prompts through iterative LLM refinement cycles.

Features

  • 🔄 Multi-Pass Refinement: Automatically runs 3-5 improvement cycles
  • 📊 Quality Scoring: Evaluates content with a 0-10 quality score
  • 🎯 Self-Correction: Identifies and fixes issues in each iteration
  • 🌐 Language Agnostic: Works with any programming language
  • 💬 Prompt Optimization: Enhances prompts for better LLM responses
  • 🔒 Privacy First: Uses your existing model, no external API calls

How It Works

┌─────────────────┐
│  User Input     │
│  (Code/Prompt)  │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Iteration 1    │
│  ┌───────────┐  │
│  │ Refactor  │──┼──► Improved v1
│  └───────────┘  │
│  ┌───────────┐  │
│  │ Evaluate  │──┼──► Score + Issues
│  └───────────┘  │
└────────┬────────┘
         │ if score < 9.0
         ▼
┌─────────────────┐
│  Iteration 2-5  │
│  (Self-Refine)  │
└────────┬────────┘
         │
         ▼
┌─────────────────┐
│  Best Version   │
│  (Final Output) │
└─────────────────┘

Usage

In VS Code Chat

@AIAutopilotEngine Optimize this function:

function calculateTotal(items) {
    var total = 0;
    for (var i = 0; i < items.length; i++) {
        total = total + items[i].price * items[i].quantity;
    }
    return total;
}

Prompt Optimization

@AIAutopilotEngine Make this prompt better:

write a function that does something with arrays

Via Command

  1. Select code in your editor
  2. Run command: AIAutopilotEngine: Start Optimization
  3. The chat will open with your selection ready to optimize

Pipeline Stages

1. Initial Generation

  • Receives your code or prompt
  • First improvement pass applied

2. Evaluation

  • Analyzes quality (0-10 score)
  • Identifies specific issues
  • Returns structured JSON feedback

3. Self-Refine Rounds

  • Targets identified issues
  • Iterates until quality threshold met
  • Maximum 5 iterations

4. Finalization

  • Returns best scoring version
  • Shows quality metrics
  • Clean output without internal details

Requirements

  • VS Code 1.101.0+
  • GitHub Copilot or compatible LLM provider

Configuration

The engine uses these default settings:

Setting Default Description
Max Iterations 6 Maximum refinement cycles
Target Score 9.5 Quality threshold to stop
Min Improvement 0.2 Minimum score gain per iteration

Privacy & Security

  • ✅ Uses your existing VS Code language model
  • ✅ No external API calls
  • ✅ No data storage
  • ✅ No API keys required
  • ✅ All processing happens locally

Supported Languages

AIAutopilotEngine works with any programming language:

  • TypeScript / JavaScript
  • Python
  • C# / .NET
  • Java
  • Go
  • Rust
  • PHP
  • Kotlin
  • And more...

License

MIT


Made with ❤️ for developers who want the best from their LLM interactions

👨‍💻 Author

Hamza Cantürk

GitHub

"Kod kalitesi tesadüf değil, sürekli iyileştirmenin sonucudur."

  • Contact us
  • Jobs
  • Privacy
  • Manage cookies
  • Terms of use
  • Trademarks
© 2025 Microsoft