Streaming output from Laminar Index API

When running browser agent tasks through the Index API, you can receive real-time updates as the agent executes rather than waiting for the entire task to complete. This is particularly useful for:

Providing immediate feedback to users
Showing progress indicators for long-running tasks
Enabling early termination if needed
Debugging agent behavior in real-time

Streaming Mode

To receive real-time updates as the browser agent executes:

import { LaminarClient } from '@lmnr-ai/lmnr';

const client = new LaminarClient({
  projectApiKey: "YOUR_PROJECT_API_KEY",
});

async function runStreamingBrowserAgent() {
  const response = await client.agent.run({
    prompt: "Lookup the current weather for New York, " +
           "London, and Tokyo and compare them.",
    stream: true,
    modelProvider: "gemini",
    model: "gemini-2.5-pro-preview-03-25",
  });
  
  for await (const chunk of response) {
    if (chunk.chunkType === 'step') {
      // Print step-by-step actions as they happen
      console.log(`Step: ${chunk.summary}`);
    } else if (chunk.chunkType === 'finalOutput') {
      // Print the final result
      console.log(`Final output: ${chunk.content.result.content}`);
    } else if (chunk.chunkType === 'timeout') {
      console.log("Final step before timeout:");
      console.log(chunk.summary);
    } else if (chunk.chunkType === 'error') {
      console.log(`Error: ${chunk.error}`);
    }
  }
}

runStreamingBrowserAgent();

To receive real-time updates as the browser agent executes:

import { LaminarClient } from '@lmnr-ai/lmnr';

const client = new LaminarClient({
  projectApiKey: "YOUR_PROJECT_API_KEY",
});

async function runStreamingBrowserAgent() {
  const response = await client.agent.run({
    prompt: "Lookup the current weather for New York, " +
           "London, and Tokyo and compare them.",
    stream: true,
    modelProvider: "gemini",
    model: "gemini-2.5-pro-preview-03-25",
  });
  
  for await (const chunk of response) {
    if (chunk.chunkType === 'step') {
      // Print step-by-step actions as they happen
      console.log(`Step: ${chunk.summary}`);
    } else if (chunk.chunkType === 'finalOutput') {
      // Print the final result
      console.log(`Final output: ${chunk.content.result.content}`);
    } else if (chunk.chunkType === 'timeout') {
      console.log("Final step before timeout:");
      console.log(chunk.summary);
    } else if (chunk.chunkType === 'error') {
      console.log(`Error: ${chunk.error}`);
    }
  }
}

runStreamingBrowserAgent();

from lmnr import LaminarClient

client = LaminarClient(project_api_key="YOUR_PROJECT_API_KEY")

# Stream the browser agent's execution
for chunk in client.agent.run(
    prompt="Lookup the current weather for New York, " +
           "London, and Tokyo and compare them.",
    stream=True,
    model_provider="gemini",
    model="gemini-2.5-pro-preview-03-25",
):
    if chunk.chunk_type == 'step':
        # Print step-by-step actions as they happen
        print(f"Step: {chunk.summary}")
    elif chunk.chunk_type == 'finalOutput':
        # Print the final result
        print(f"Final output: {chunk.content.result.content}")
    elif chunk.chunk_type == 'timeout':
        print("Final step before timeout:")
        print(chunk.summary)
    elif chunk.chunk_type == 'error':
        print(f"Error: {chunk.error}")

import asyncio
from lmnr import AsyncLaminarClient

async def run_streaming_agent():
    client = AsyncLaminarClient(project_api_key="YOUR_PROJECT_API_KEY")
    
    # Stream the browser agent's execution asynchronously
    async for chunk in await client.agent.run(
        prompt="Lookup the current weather for New York, " +
               "London, and Tokyo and compare them.",
        stream=True,
        model_provider="gemini",
        model="gemini-2.5-pro-preview-03-25",
    ):
        if chunk.chunk_type == 'step':
            print(f"Step: {chunk.summary}")
        elif chunk.chunk_type == 'finalOutput':
            print(f"Final output: {chunk.content.result.content}")
        elif chunk.chunk_type == 'timeout':
            print("Final step before timeout:")
            print(chunk.summary)
        elif chunk.chunk_type == 'error':
            print(f"Error: {chunk.error}")

# Run the async function
asyncio.run(run_streaming_agent())

Chunk Types

When streaming browser agent output, you’ll receive different types of chunks that represent various aspects of the agent’s execution:

Chunk Type	Description
`step`	A step taken by the agent during execution (e.g., navigating to a page, clicking a button)
`finalOutput`	The final result of the agent’s execution
`timeout`	The agent timed out. It’s no different from a step chunk, but this will be the last chunk in the stream.
`error`	The agent encountered an error. The error message is in the `error` field.

Index

Serverless API

​Streaming Mode

​Chunk Types

Streaming Mode

Chunk Types