WebBrain Launches as the Free, Open-Source AI Browser Agent for Chrome and Firefox

The privacy-first, multi-provider alternative to proprietary AI browser plugins is now available on the Chrome Web Store and Firefox Add-ons

April 28, 2026 — Developer Emre Sokullu today announced the public launch of WebBrain, a free, open-source browser extension that brings AI agent capabilities to Chrome and Firefox. WebBrain lets users read pages, extract structured data, and automate multi-step web tasks — all powered by their choice of large language model, including fully offline local models.

Unlike proprietary AI browser plugins that lock users into a single provider and require paid subscriptions, WebBrain supports seven LLM backends out of the box: llama.cpp, Ollama, OpenAI, Claude, OpenRouter, StudioLM, and vLLM. Users can run the extension entirely offline with a local model — no API keys, no cloud dependency, and no data ever leaving their machine.

A Browser AI That Puts Users in Control

WebBrain operates in two modes. Ask mode (the default) is read-only: users can ask questions about any web page, extract information, and summarize content without the extension modifying anything. Act mode unlocks full browser agent capabilities — clicking buttons, filling forms, navigating between pages, and automating complex workflows — with explicit user confirmation and a visible warning banner before any action is taken.

"I built WebBrain because I wanted an AI browser assistant that didn't force me into one provider or require a subscription," said Emre Sokullu, creator of WebBrain. "The browser is where most of us spend our working day. Having an AI agent that understands any page you're on, works with any model you prefer, and can run completely on your own hardware — that should be free and open to everyone."

Key Features

Multi-provider LLM support: Connect to llama.cpp, Ollama, OpenAI, Claude, OpenRouter, StudioLM, vLLM, or any OpenAI-compatible API endpoint
Fully offline operation: Run with a local GGUF model via llama.cpp for zero data leakage
Cross-browser: Available for Chrome (Manifest V3) and Firefox (Manifest V2) with identical feature sets
Token-conscious design: Screenshots are automatically resized and compressed before being sent to any model, keeping cloud API costs predictable
Dedicated vision model: Pair a fast text model for planning with a separate vision model for screenshots, reducing costs without sacrificing capability
Multilingual UI: Ships in English, Spanish, French, Turkish, and Chinese with automatic browser language detection
Cookie and paywall awareness: Automatically dismisses consent banners and honestly reports paywalls instead of fabricating content
Tab groups: Agent-opened tabs are organized into a dedicated group to keep research chains together
Profile auto-fill: Optional feature for low-stakes signup forms, stored locally and off by default
Safety-first agent design: Act mode requires explicit confirmation; all mutations go through the visible UI rather than hidden API calls

Availability

WebBrain is available now at no cost:

Chrome Web Store: https://chromewebstore.google.com/detail/webbrain/ljhijonmfahplgbbacgcfnaihbjljhhb
Firefox Add-ons: https://addons.mozilla.org/en-US/firefox/addon/webbrain/
Source code (MIT License): https://github.com/esokullu/webbrain
Website: https://www.webbrain.one

About the Creator

Emre Sokullu is an independent developer focused on open-source AI tools. WebBrain is MIT-licensed and welcomes contributions from the community.

Media Contact:
Emre Sokullu
Email: esokullu@gmail.com
Website: https://emresokullu.com

###