Skip to main content
This guide shows you how to add observability to your Node.js/TypeScript LLM application.

​
Prerequisites

  • Muxx TypeScript SDK installed (npm install muxx)
  • Your Muxx API key
  • An LLM provider SDK (OpenAI, Anthropic, etc.)

​
Basic Usage

​
Wrap Your Client

The simplest way to use Muxx is to wrap your LLM client: 
import { Muxx } from 'muxx';
import OpenAI from 'openai';

// Initialize Muxx
const muxx = new Muxx({ apiKey: 'muxx_sk_live_xxxxxxxxxxxx' });

// Wrap the OpenAI client
const client = muxx.wrap(new OpenAI());

// Use as normal - all calls are automatically traced
const response = await client.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'What is the capital of France?' }],
});

console.log(response.choices[0].message.content);
That’s it! All requests are now logged to Muxx.

​
Adding Context

​
Traces

Group related operations into traces: 
import { Muxx } from 'muxx';
import OpenAI from 'openai';

const muxx = new Muxx();
const client = muxx.wrap(new OpenAI());

async function summarizeDocument(document: string): Promise<string> {
  return muxx.trace('document-summary', async () => {
    // Extract key points
    const points = await client.chat.completions.create({
      model: 'gpt-4o',
      messages: [{ role: 'user', content: `Extract key points from: ${document}` }],
    });

    // Generate summary
    const summary = await client.chat.completions.create({
      model: 'gpt-4o',
      messages: [
        { role: 'user', content: `Summarize these points: ${points.choices[0].message.content}` },
      ],
    });

    return summary.choices[0].message.content ?? '';
  });
}

// Both LLM calls are grouped under one trace
const result = await summarizeDocument('Your long document here...');

​
Spans

Add more granular tracking with spans: 
import { Muxx } from 'muxx';
import OpenAI from 'openai';

const muxx = new Muxx();
const client = muxx.wrap(new OpenAI());

async function summarizeDocument(document: string): Promise<string> {
  return muxx.trace('document-summary', async () => {
    const points = await muxx.span('extract-points', async () => {
      return client.chat.completions.create({
        model: 'gpt-4o',
        messages: [{ role: 'user', content: `Extract key points: ${document}` }],
      });
    });

    const summary = await muxx.span('generate-summary', async () => {
      return client.chat.completions.create({
        model: 'gpt-4o',
        messages: [{ role: 'user', content: `Summarize: ${points.choices[0].message.content}` }],
      });
    });

    return summary.choices[0].message.content ?? '';
  });
}

​
Metadata

Add custom metadata for filtering: 
await muxx.trace(
  'user-chat',
  async () => {
    // Your code here
  },
  { metadata: { userId: 'user_123', feature: 'support' } }
);

​
Error Handling

Errors are automatically captured in traces: 
await muxx.trace('risky-operation', async () => {
  try {
    await riskyOperation();
  } catch (error) {
    // Error is captured with full stack trace
    throw error;
  }
});

​
Next Steps

Tracing

Learn more about traces and spans

OpenAI

OpenAI-specific features