A few weeks ago I decided to build something that could actually help me (and other devs) catch bugs and bad patterns before they ship — so I built CodeMate, a free VS Code extension for instant AI code review.

What it does

Select any code, run the command, and it gives you:

• A code score out of 10
• Bugs, categorized by severity (Critical/High/Medium/Low)
• Performance issues
• Security vulnerabilities
• A suggested "fixed" version — but it doesn't auto-apply changes. The idea is you read the explanation and fix it yourself, so you actually learn from the mistake instead of just accepting a diff blindly.

How it works

Under the hood it uses Groq's API (Llama 3.3 70B) for fast, free inference. Each user provides their own free Groq API key, so there's no cost on my end and no shared rate limits.

Why I built it

I wanted something lightweight that doesn't require a subscription or a heavy setup, and that nudges people toward understanding their bugs rather than just auto-fixing everything for them.

Current state

This is still early (v0.5), built solo over a couple of weeks. It's rough in places, and I'd love feedback on what's actually useful vs what's missing.

🔗 Try it on the VS Code Marketplace

Happy to answer any questions about how it's built or what's next!