> ## Documentation Index
> Fetch the complete documentation index at: https://docs.keywordsai.co/llms.txt
> Use this file to discover all available pages before exploring further.

# initialize()

> Initialize the Keywords AI telemetry client

## Overview

The `initialize()` method sets up the tracing client and establishes connection to Keywords AI. This must be called before using any tracing methods.

## Signature

```typescript theme={"system"}
async initialize(): Promise<void>
```

## Basic Usage

```typescript theme={"system"}
import { KeywordsAITelemetry } from '@keywordsai/tracing';

const keywordsAi = new KeywordsAITelemetry({
    apiKey: process.env.KEYWORDSAI_API_KEY,
    appName: 'my-app'
});

// Initialize before using any tracing methods
await keywordsAi.initialize();

// Now ready to use
await keywordsAi.withWorkflow(
    { name: 'my_workflow' },
    async () => {
        return 'ready';
    }
);
```

## Constructor Options

```typescript theme={"system"}
const keywordsAi = new KeywordsAITelemetry({
    apiKey: string;              // Required: Your Keywords AI API key
    baseURL?: string;            // Optional: API base URL (default: https://api.keywordsai.co)
    appName?: string;            // Optional: Application name for identification
    instrumentModules?: {        // Optional: Modules to auto-instrument
        openAI?: typeof OpenAI;
        anthropic?: typeof Anthropic;
    };
    disableBatch?: boolean;      // Optional: Send spans immediately (default: false)
    logLevel?: string;           // Optional: 'debug' | 'info' | 'warn' | 'error' (default: 'warn')
});
```

## Complete Example

```typescript theme={"system"}
import { KeywordsAITelemetry } from '@keywordsai/tracing';
import OpenAI from 'openai';

const keywordsAi = new KeywordsAITelemetry({
    apiKey: process.env.KEYWORDSAI_API_KEY,
    baseURL: 'https://api.keywordsai.co',
    appName: 'production-app',
    instrumentModules: {
        openAI: OpenAI,  // Auto-instrument OpenAI
    },
    disableBatch: false,
    logLevel: 'info'
});

// Initialize the client
await keywordsAi.initialize();
console.log('Keywords AI Tracing initialized successfully');

// Use tracing
await keywordsAi.withWorkflow(
    { name: 'user_request' },
    async () => {
        // Your code here
    }
);

// Shutdown gracefully when done
await keywordsAi.shutdown();
```

## Environment Variables

You can also configure using environment variables:

```bash .env theme={"system"}
KEYWORDSAI_API_KEY=your-api-key
KEYWORDSAI_BASE_URL=https://api.keywordsai.co
```

```typescript theme={"system"}
const keywordsAi = new KeywordsAITelemetry({
    apiKey: process.env.KEYWORDSAI_API_KEY,
    baseURL: process.env.KEYWORDSAI_BASE_URL,
    appName: 'my-app'
});

await keywordsAi.initialize();
```

## Configuration Options

<ParamField body="apiKey" type="string" required>
  Your Keywords AI API key from the dashboard
</ParamField>

<ParamField body="baseURL" type="string" default="https://api.keywordsai.co">
  Keywords AI API base URL
</ParamField>

<ParamField body="appName" type="string">
  Application name for identifying traces in the dashboard
</ParamField>

<ParamField body="instrumentModules" type="object">
  Modules to automatically instrument:

  * `openAI`: OpenAI SDK class
  * `anthropic`: Anthropic SDK class
</ParamField>

<ParamField body="disableBatch" type="boolean" default={false}>
  If true, sends spans immediately instead of batching them
</ParamField>

<ParamField body="logLevel" type="string" default="warn">
  Logging verbosity: `"debug"`, `"info"`, `"warn"`, `"error"`
</ParamField>

## Best Practices

* Always call `initialize()` before using any tracing methods
* Initialize once at application startup
* Use environment variables for sensitive configuration
* Call `shutdown()` before application exit to flush pending spans
* Enable auto-instrumentation for supported SDKs