🤖 AI Error Explainer

AI-powered error explanations, right inside VS Code.

Hover over any error or warning to get a clear, plain-language diagnosis of what went wrong, why it happened, and how to fix it — powered by the AI of your choice.

📑 Table of Contents

• Demo
• Why This Extension?
• Features
• Supported AI Providers
• Example
• Setup
• How It Works
• Settings Reference
• Privacy & Security
• FAQ
• Known Limitations
• License
• Author

🎬 AI Error Explainer Example

https://github.com/user-attachments/assets/c96ee44a-e554-4d8e-9f57-f70f31548828

💡 Why This Extension?

Cryptic error messages waste time. Instead of copy-pasting errors into a browser tab and searching through forum posts, get an instant, plain-language diagnosis without ever leaving your editor.

This extension follows one core principle: bring your own key. There's no subscription, no server in the middle, and no vendor lock-in — you choose the AI provider you already trust, and every request goes directly from your machine to theirs.

✨ Features

🔌 Supported AI Providers

This extension works with any AI provider — bring the key you already have. Choose your setup path below based on which provider you're using.

Api Format: gemini
   or
Api Format: anthropic

Api Key: <your key>

Api Format: openai-compatible

Base URL: <provider's endpoint>
Model:    <provider's model name>
Api Key:  <your key>

Provider Directory

💡 Not seeing your provider? Any service offering an OpenAI-compatible chat completions endpoint will likely work — just fill in its Base URL and Model name the same way as above.

How It Works Internally

┌─────────────────────────────────────────────┐
│  Api Format: gemini / anthropic              │
│  → Extension already knows the correct URL   │
│  → You only provide: API Key                 │
├─────────────────────────────────────────────┤
│  Api Format: openai-compatible               │
│  → Hundreds of providers share this format   │
│  → You provide: Base URL + Model + API Key   │
└─────────────────────────────────────────────┘

📸 Example

Before — you hit a confusing error:

const users = getUsers();
users.map(user => console.log(user.name));

TypeError: Cannot read properties of undefined (reading 'map')

Hover over the red underline, and AI Error Explainer instantly shows:

🤖 AI Error Assistant

What happened? The users variable is undefined. Why? getUsers() did not return an array. Fix: Make sure getUsers() returns an array, or add a fallback like getUsers() || [].

🔍 View Details

Click View Details to open a dedicated side panel with a clean, full breakdown — no squinting at a tiny tooltip.

⚙️ Setup

1. Install the extension from the VS Code Marketplace
2. Pick a provider from the Supported AI Providers section above and grab a free API key
3. Open VS Code Settings (Ctrl+,) and search for "AI Error Explainer"
4. Follow either the Quick Setup (Gemini/Claude) or Advanced Setup (everything else) instructions
5. (Optional) Set your preferred Language
6. That's it — hover over any error or warning in your code

🔄 How It Works

Your code
   ↓
VS Code detects an error
   ↓
You hover over it
   ↓
Extension checks its cache — if this exact error was seen before,
the cached answer is shown instantly
   ↓
Otherwise: sends the error message + a few nearby lines to your chosen AI
   ↓
AI explains what happened, why, and how to fix it
   ↓
Click "View Details" for the full diagnosis

Only a small snippet — the error message and a few surrounding lines — is ever shared. Never your whole file, never your whole project.

🛠️ Settings Reference

🔐 Privacy & Security

• ✅ No data is collected or stored by this extension. Nothing is sent to any server we control — there is no "we" in the data path at all.
• ✅ Your API key never leaves your machine except to talk directly to the provider you chose. It's stored in VS Code's local settings, same as any other extension setting.
• ✅ Only a small snippet is shared — the error message and a few nearby lines, never your whole file.
• ✅ This extension never modifies your code. It only explains errors — it does not apply or suggest automatic changes to your files.
• ✅ Caching is local and temporary. Cached responses live only in memory for your current VS Code session and are cleared on restart — nothing is written to disk.
• ✅ No telemetry, no analytics, no tracking.

❓ FAQ

Is this free to use? The extension itself is completely free. Whether using it is free depends on your chosen AI provider — Gemini, OpenRouter, and Groq all offer generous free tiers with no billing setup required.

Do I need to pay you anything? No. There is no subscription, no license key, and no payment to us at all. You pay (or don't pay) your AI provider directly.

Which provider should I pick if I have no key yet? Start with Google Gemini (Quick Setup, generous free tier) or Groq (Advanced Setup, extremely fast and free) — both take under a minute to set up.

What's the difference between "Quick Setup" and "Advanced Setup"? Gemini and Claude use their own unique request format, so the extension already knows how to talk to them — you just paste a key. Every other provider (OpenAI, DeepSeek, Groq, Mistral, OpenRouter, Together AI, xAI) shares a common "OpenAI-compatible" format, so you tell the extension the provider's Base URL and Model name once.

Why isn't there an "Apply Fix" button that edits my code? Early versions had one, but auto-replacing multiple lines of code carries real risk of breaking working code — especially in larger files. We'd rather give you a clear, trustworthy explanation than a risky auto-edit.

My AI provider says "rate limit exceeded" or asks for billing — is this extension broken? No — that message comes directly from your provider's account status, not from a bug in the extension. Wait for a free-tier limit to reset, add billing on that provider's dashboard, or switch to a different provider in Settings.

⚠️ Known Limitations

• Free-tier AI models may respond more slowly or hit rate limits depending on the provider and time of day
• Very large code blocks are truncated before being sent to the AI to keep requests fast and affordable
• This extension only explains errors — it does not automatically modify your code
• Some providers change their model names over time; if a model stops working, check that provider's documentation for the current name

📄 License

MIT — free to use, modify, and share. See the LICENSE file included with this extension for details.

👤 Author

Built by Robin Ryan, a full-stack developer from Bangladesh.

If you run into an issue or have a feature request, feel free to leave a review or rating on the Marketplace listing — feedback helps shape future updates.