The SDK provides two clients. InfiniumClient is the standard choice — all its methods already return Promises. AsyncInfiniumClient adds explicit lifecycle management with close() and additional batch methods.

When to Use Which

Feature	InfiniumClient	AsyncInfiniumClient
sendTask()	Promise<ApiResponse>	Promise<ApiResponse>
sendBatch()	Sequential	Concurrent
sendBatchSequential()	N/A	Sequential
close()	No-op	Cleanup resources
isClosed()	N/A	Check state

Use AsyncInfiniumClient when you need concurrent batch processing, explicit resource cleanup, or graceful shutdown handling.

AsyncInfiniumClient

import { AsyncInfiniumClient } from 'infinium-o2';

const client = new AsyncInfiniumClient({
  agentId: 'your-agent-id',
  agentSecret: 'your-agent-secret',
});

try {
  await client.sendTask({
    name: 'Moderate content',
    description: 'Check user content for policy violations.',
    duration: 1.8,
  });
} finally {
  await client.close();
}

Lifecycle Methods

close(): Promise<void> — Clean up resources. After calling, further requests will fail.

isClosed(): boolean — Check if the client has been closed.

Concurrent Batch

AsyncInfiniumClient.sendBatch() processes tasks concurrently:

const result = await client.sendBatch(tasks);
console.log(`Sent ${result.successfulTasks}/${result.totalTasks} traces`);

For strict rate limiting, use sequential processing:

const result = await client.sendBatchSequential(tasks);

Trace Wrapper

Works identically to InfiniumClient:

const classify = client.trace('Classifier')(
  async (text: string) => {
    const resp = await openai.chat.completions.create({
      model: 'gpt-4o',
      messages: [{ role: 'user', content: text }],
    });
    return JSON.parse(resp.choices[0].message.content!);
  }
);

With watch()

import OpenAI from 'openai';
import Anthropic from '@anthropic-ai/sdk';
import { AsyncInfiniumClient, watch } from 'infinium-o2';

const client = new AsyncInfiniumClient({ agentId: '...', agentSecret: '...' });
const openai = watch(new OpenAI());
const anthropic = watch(new Anthropic());

const process_ = client.trace('Multi-Provider Agent')(
  async (query: string) => {
    const resp1 = await openai.chat.completions.create({
      model: 'gpt-4o',
      messages: [{ role: 'user', content: query }],
    });
    const resp2 = await anthropic.messages.create({
      model: 'claude-3-5-sonnet-20241022',
      max_tokens: 1024,
      messages: [{ role: 'user', content: query }],
    });
    return { openai: resp1.choices[0].message.content, anthropic: resp2.content[0].text };
  }
);

Graceful Shutdown

Handle process signals for clean shutdown:

const client = new AsyncInfiniumClient({ agentId: '...', agentSecret: '...' });

async function shutdown() {
  console.log('Shutting down...');
  await client.close();
  process.exit(0);
}

process.on('SIGTERM', shutdown);
process.on('SIGINT', shutdown);

Maestro Polling

const response = await client.sendTaskData(taskData);
const traceId = response.data?.traceId;

const interpretation = await client.getInterpretedTaskResult(traceId);
console.log(interpretation.data);

Prompt Studio

const prompt = await client.getPrompt(
  '...', '...',
  'latest',
  { tone: 'professional' },
);