Index: The SOTA Open-Source Browser Agent for Autonomous Web Tasks

Introduction

Index, developed by lmnr-ai, is a state-of-the-art open-source browser agent that empowers users to autonomously perform complex tasks on the web. It effectively turns any website into an accessible API, allowing for seamless integration into your projects with just a few lines of code. Index leverages powerful reasoning Large Language Models (LLMs) with vision capabilities, supporting models like Gemini 2.5 Pro, Claude 3.7 Sonnet, and OpenAI o4-mini, to navigate, interact with, and extract information from web pages intelligently.

Installation

Getting started with Index is straightforward. You can install it using pip and set up the necessary browser dependencies.

pip install lmnr-index 'lmnr[all]'

# Install playwright
playwright install chromium

After installation, remember to set up your model API keys (GEMINI_API_KEY, ANTHROPIC_API_KEY, OPENAI_API_KEY) in a .env file in your project root. For advanced observability, you can also include your LMNR_PROJECT_API_KEY.

Examples

Index offers flexible ways to interact with the agent, whether through code, an interactive CLI, or a serverless API.

Run Index with Code

You can integrate Index directly into your Python applications to perform tasks and extract structured data using Pydantic schemas.

import asyncio
from index import Agent, GeminiProvider
from pydantic import BaseModel
from lmnr import Laminar
import os

# to trace the agent's actions and record browser session
Laminar.initialize()

# Define Pydantic schema for structured output
class NewsSummary(BaseModel):
    title: str
    summary: str

async def main():
    llm = GeminiProvider(model="gemini-2.5-pro-preview-05-06")
    agent = Agent(llm=llm)

    # Example of getting structured output
    output = await agent.run(
        prompt="Navigate to news.ycombinator.com, find a post about AI, extract its title and provide a concise summary.",
        output_model=NewsSummary
    )
    
    summary = NewsSummary.model_validate(output.result.content)
    print(f"Title: {summary.title}")
    print(f"Summary: {summary.summary}")
    
if __name__ == "__main__":
    asyncio.run(main())

Run Index with CLI

For interactive use and quick tasks, Index provides a powerful command-line interface with features like browser state persistence, follow-up messages, and real-time streaming updates.

index run

You can also run the CLI with your personal Chrome instance to leverage existing logged-in sessions:

index run --local-chrome

Use Index via API

For production environments, Index is available as a serverless API, managing remote browser sessions, agent infrastructure, and browser observability.

from lmnr import Laminar, LaminarClient

# Initialize tracing (optional, but recommended for observability)
Laminar.initialize(project_api_key="your_api_key")

# Initialize the client
client = LaminarClient(project_api_key="your_api_key")

for chunk in client.agent.run(
    stream=True,
    model_provider="gemini",
    model="gemini-2.5-pro-preview-05-06",
    prompt="Navigate to news.ycombinator.com, find a post about AI, and summarize it"
):
    print(chunk)

Why Use Index?

Index stands out as a robust solution for web automation due to several key advantages:

State-of-the-Art Performance: It's built to be a leading open-source browser agent, capable of handling complex web tasks autonomously.
Flexible LLM Integration: Supports multiple powerful LLMs, including Gemini, Claude, and OpenAI models, allowing you to choose the best fit for your needs in terms of speed, cost, and accuracy.
Structured Output: Enables reliable data extraction by supporting Pydantic schemas, ensuring your extracted data is clean and usable.
Advanced Observability: Integrates with Laminar for comprehensive tracing of agent actions and browser sessions, providing deep insights into its operations.
Ease of Use: With simple pip install and intuitive CLI and API options, Index is designed for quick deployment and integration.

Index: The SOTA Open-Source Browser Agent for Autonomous Web Tasks

Summary

Repository Info

Tags