ai-log-debugger-sdk

Official Node.js SDK for AI Log Debugger — automatically captures uncaught errors and unhandled promise rejections from your Node.js application and reports them to your AI Log Debugger dashboard with AI-powered fix suggestions.

No manual error handling required. Install, initialize with 2 lines, and you're done.

How It Works

Your Node.js App
      │
      │  uncaughtException / unhandledRejection
      ▼
ai-log-debugger-sdk
      │
      │  POST /api/errors/report
      ▼
AI Log Debugger Dashboard
      │
      │  AI analyzes error → suggests fix
      ▼
You get notified (Slack / Email)

Requirements

• Node.js 16 or higher
• An AI Log Debugger account
• Your API token (available in your dashboard after signup)

Installation

npm install ai-log-debugger-sdk

Quick Start

Add these 2 lines at the very top of your main file (e.g. server.js, app.js, or index.js):

const aiDebugger = require('ai-log-debugger-sdk');

aiDebugger.init({
  apiToken: 'your-api-token',   // from your AI Log Debugger dashboard
  endpoint: 'https://your-backend.com/api/errors/report',
});

That's it. Every uncaught error and unhandled promise rejection in your app will now be automatically captured and reported to your dashboard.

Getting Your API Token

1. Sign up at AI Log Debugger
2. Go to Settings → API Token
3. Copy your token and paste it into init()

Usage Examples

Basic Node.js App

const aiDebugger = require('ai-log-debugger-sdk');

aiDebugger.init({
  apiToken: 'your-api-token',
  endpoint: 'https://your-backend.com/api/errors/report',
});

// From this point, all uncaught errors are automatically reported.
// Your app continues to work as normal.

Express.js App

const express = require('express');
const aiDebugger = require('ai-log-debugger-sdk');

const app = express();

// Initialize at the top — before your routes
aiDebugger.init({
  apiToken: 'your-api-token',
  endpoint: 'https://your-backend.com/api/errors/report',
});

// Your routes
app.get('/', (req, res) => {
  res.send('Hello World');
});

// Add error middleware at the very end — after all routes
app.use(aiDebugger.expressErrorHandler());

app.listen(3000);

Manual Error Capture

Capture specific errors inside try/catch blocks:

try {
  await riskyOperation();
} catch (err) {
  // Report with default severity (MEDIUM)
  aiDebugger.captureError(err);

  // Or specify severity manually
  aiDebugger.captureError(err, { severity: 'CRITICAL' });
}

API Reference

aiDebugger.init(options)

Initializes the SDK and attaches global error hooks.

| Option | Type | Required | Description | |--------|------|----------|-------------| | apiToken | string | ✅ Yes | Your API token from the dashboard | | endpoint | string | No | Your backend URL (defaults to the AI Log Debugger endpoint) |

aiDebugger.captureError(err, options?)

Manually capture and report any error.

| Parameter | Type | Description | |-----------|------|-------------| | err | Error | The error to capture | | options.severity | string | LOW, MEDIUM, HIGH, or CRITICAL (default: MEDIUM) |

Returns a Promise<void>.

aiDebugger.expressErrorHandler()

Returns an Express error-handling middleware. Must be placed after all routes.

app.use(aiDebugger.expressErrorHandler());

Severity Levels

| Level | When to use | |-------|-------------| | LOW | Minor issues, non-critical warnings | | MEDIUM | Standard errors (default) | | HIGH | Uncaught exceptions, serious failures | | CRITICAL | Fatal crashes, data loss risk |

Severity values are case-insensitive — critical, Critical, and CRITICAL all work.

Default Severity by Error Type

| Error Type | Default Severity | |------------|-----------------| | uncaughtException | HIGH | | unhandledRejection | MEDIUM | | Express route error | HIGH | | Manual captureError() | MEDIUM |

Error Reporting Payload

Each error is sent to your backend with this structure:

{
  "api_token": "your-token",
  "error_text": "Cannot read properties of undefined (reading 'id')",
  "ai_fix": "Analyzing...",
  "severity": "HIGH"
}

The ai_fix field is populated by your AI Log Debugger backend — the SDK itself does not call any AI APIs.

Reliability

• Non-blocking — the SDK never slows down or crashes your app
• Retry on failure — automatically retries once if the network request fails
• Silent failure — if reporting fails after retry, a warning is logged but your app continues normally
• No dependencies — uses only Node.js built-in modules (http, https)

Testing

npm test

Uses Node.js built-in test runner — no extra dependencies required.

Test coverage includes:

• init() configuration and hook setup
• Correct payload sent to backend (via mock HTTP server)
• uncaughtException and unhandledRejection hook behavior
• Severity auto-detection and normalization
• Network failure handling (no crash on failure)
• Express middleware behavior

Coming Soon

• 🐍 Python SDK — pip install ai-log-debugger-sdk
• 🐘 PHP SDK — composer require ai-log-debugger/sdk
• ☕ Java SDK
• 🟦 TypeScript types for this package

License

MIT — see LICENSE for details.

Links

• Dashboard
• Report an Issue
• npm Package