webscrape.dev

Open-source LLM agent framework for browser automation

Nathan Kessler
By Nathan KesslerUpdated

Each tool is evaluated against our methodology using public docs, vendor demos, and hands-on testing.

Some links on this page are affiliate links. We earn a commission if you sign up – at no additional cost to you. Our editorial assessment is independent and never paid. How we review.

Agent-E website

What is Agent-E?

Agent-E is an open-source LLM agent framework for browser automation from Emergence AI, built on the AG2 (formerly AutoGen) framework. It uses a skills-and-agents architecture based on DOM parsing, without a vision model, to carry out natural-language browser tasks such as filling web forms, searching and sorting e-commerce listings, and running web searches. The project ships a FastAPI wrapper with an /execute_task endpoint for programmatic HTTP access and streaming results, and is available on GitHub under an MIT license with around 1.2K stars.

Our verdict

Agent-E's DOM-only approach avoids vision-model overhead but trails Browser Use and Skyvern 2.0 on the WebVoyager benchmark (73.1% versus 89.1% and 85.8% respectively). It suits teams that want an open-source, self-hostable framework to build on, though Emergence AI's own site now centers on a broader enterprise platform rather than the open-source project.

Categories:

Agentic automation tools give an LLM control of a browser to navigate, click, and fill forms toward a goal, rather than following a fixed script. They are newer and less predictable than selector-based scraping, but they handle sites and workflows that are hard to hardcode.

Share:

Open-source browser agent framework plus a paid cloud API

FreemiumJul 2026Agentic Automation

Open source browser automation agent using LLMs and vision

FreemiumJul 2026Agentic Automation

Open-source framework for building AI agents to navigate websites

FreemiumJul 2026Agentic Automation

How Agent-E compares

Browser Use

Browser Use is another open-source browser automation framework for LLM agents and outscored Agent-E on the WebVoyager benchmark, 89.1% versus 73.1%.

Skyvern

Skyvern 2.0 combines vision and DOM-based automation and also outperformed Agent-E on WebVoyager, 85.8% versus 73.1%.

LaVague

LaVague, also open-source, is an AI agent framework for browser automation and offers a similar self-hosted alternative to Agent-E.

Visit

Agent-E

Visit