Repolode
Submit
webbrain-one

webbrain-one/webbrain

Open-source AI browser agent extension for Chrome, Firefox and Edge that chats with pages and automates multi-step browser tasks using your choice of LLM.

View on GitHubWebsite
Stars
1,668
Forks
209
Open issues
6
Last commit
14 hours ago

What is it?

WebBrain is a browser extension that places an AI agent in a side panel beside your tabs. You can ask questions about the current page, or give it a task and let it click, type, fill forms, and navigate through the site. It works with a built-in default model that needs no setup, with local servers such as llama.cpp, Ollama, vLLM, LM Studio, and with a long list of cloud LLM providers including OpenAI, Anthropic, Google Gemini, and OpenRouter.

The agent operates in three modes: Ask is read-only, Act can click, type, navigate, and upload or download files, and Dev adds page source, styles, console, and network inspection. Each mode is gated by per-site permission prompts for consequential actions. Pages are read through the accessibility tree, and a vision mode can work from screenshots with coordinate-based clicks. Users can also save successful runs as reusable workflows, schedule tasks, and set watches that poll a page for a condition.

The project also includes an MCP server so coding agents such as Claude Code, Codex, and Cursor can delegate browser tasks to WebBrain running in the user's signed-in browser, and an LM Studio plugin. Firefox support is described as weaker than Chrome because it lacks the Chrome DevTools Protocol.

Key features

  • 01Chat with and ask questions about the current page, including text, links, forms, tables, and PDFs.
  • 02Click, type, scroll, navigate, upload, and download with per-site permission prompts.
  • 03Three modes (Ask, Act, Dev) control what the agent is allowed to do.
  • 04Works with local OpenAI-compatible servers and many cloud LLM providers.
  • 05Saved workflows, scheduled tasks, and page watches for repeatable jobs.
  • 06MCP server lets coding agents delegate browser tasks to your signed-in browser.

Use cases

  • Summarizing articles or finding specific information on a web page without leaving it.
  • Automating repetitive form filling or multi-step web workflows in your own browser session.
  • Running a local model for browser automation while keeping page content on your machine.
  • Letting a coding agent use a signed-in browser for tasks behind login walls.

Quick start

  1. 1

    Install the extension from the Chrome Web Store

  2. 2

    Click the WebBrain icon to open the side panel and start chatting; the default model needs no setup

Things to know

  • The default model is a managed service, and local or cloud models require an API key or a separate local server.
  • The license is listed as NOASSERTION in metadata, though the README links GPL-3.0-or-later.

In collections

Browser automation for AI agentsOpen-source AI agent frameworks & appsMCP serversChrome & browser extensions

Topics

#ai-agent#browser-automation#browser-extension#chrome-extension#firefox-addon#llm#offline-llm#mcp#ai-agents#ai-tools

FAQ

What is webbrain?
Open-source AI browser agent extension for Chrome, Firefox and Edge that chats with pages and automates multi-step browser tasks using your choice of LLM. WebBrain is a browser extension that places an AI agent in a side panel beside your tabs. You can ask questions about the current page, or give it a task and let it click, type, fill forms, and navigate through the site. It works with a built-in default model that needs no setup, with local servers such as llama.cpp, Ollama, vLLM, LM Studio, and with a long list of cloud LLM providers including OpenAI, Anthropic, Google Gemini, and OpenRouter.
Is webbrain free to use? What license does it have?
webbrain is open source under the Other license. Check the license terms before using it in commercial products.
Is webbrain still maintained?
Actively maintained. The last commit to webbrain was on Nov 6, 2026.
What can I use webbrain for?
Summarizing articles or finding specific information on a web page without leaving it. Automating repetitive form filling or multi-step web workflows in your own browser session. Running a local model for browser automation while keeping page content on your machine. Letting a coding agent use a signed-in browser for tasks behind login walls.
How do I install webbrain?
Install the extension from the Chrome Web Store Click the WebBrain icon to open the side panel and start chatting; the default model needs no setup
Where can I get webbrain?
The source code and installation instructions are on GitHub: https://github.com/webbrain-one/webbrain

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects