OpenBrowser

OpenBrowser is a self-hosted browser automation agent for people who need visual judgment and interaction across web pages. Its server and Chrome extension run locally, but the documented Qwen setup sends model requests to DashScope's cloud API and requires an API key and internet access. Browser execution stays local; model processing uses the remote provider.

The agent works with screenshots and browser actions, so tasks can include interpreting images as well as collecting text. Its Zillow demonstration compares rental photos and listing details before returning a shortlist with reasons and links. It can also extract website data, fill and submit forms, and manage tabs while preserving cookies and login state across tasks. Chrome is the supported browser.

For recurring tasks, OpenBrowser turns a manual demonstration into a saved Browser Routine. You can review captured actions and screenshots before the agent interprets the workflow. It asks for clarification when the recording is ambiguous, and later runs follow the interpreted routine rather than reproduce every recorded click.

A web interface supports direct tasks, while an API and CLI provide programmatic access. Included skills connect browser control to Claude Code, Codex and OpenClaw. Browser execution runs separately from the control window to keep the full browser history out of the controlling model's context.

OpenBrowser is open source under LGPL-3.0 and builds on the OpenHands SDK. Its documented model choices include Qwen3.5 Plus and Flash, with evaluation coverage for Qwen3.6 models.