Open Browser Use
Open-source browser automation for local AI agents
110 followers
Open-source browser automation for local AI agents
110 followers
Open Browser Use connects local AI agents to your real Chrome profile through an open-source MV3 extension, native host, CLI, MCP server, and JS/Python/Go SDKs. It can open and claim tabs, run CDP commands, inspect page state, watch downloads, handle file choosers, and keep agent tabs organized without a hosted automation service.
Free
Launch Team
Viktor.comAn AI coworker that actually does the work
Promoted
Open Browser Use
Warmup Inbox
@ifuryst Do you support proxy management?
Open Browser Use
@fabian_maume Not built in as a dedicated layer yet. Right now OBU follows whatever proxy setup Chrome/the OS is using.
What do you mean by proxy management specifically: per-profile config, per-session proxy, auth proxies, or proxy rotation?
Warmup Inbox
@ifuryst proxy rotation. When you run webscraping localy it is something you will need quite fast.
Open-source browser use is the missing link for local agents. Does this support standard MCP (Model Context Protocol), or is it a custom automation framework?
Open Browser Use
@rivra_dev Yes - it supports standard MCP. OBU includes an MCP server, plus CLI and JS/Python/Go SDKs if you want to use it outside an MCP client.
So it is not a custom-only framework; MCP is one of the main integration paths.
Running against your actual Chrome profile is smart - sites behave very differently with real history and cookies. How do you handle CDP detection? Thats where most local automation setups get flaky.
Open Browser Use
@christian_knaut Thanks! OBU communicates through a Chrome extension plus a native host binary. There is a thin wrapper in the middle, mostly to make the surface easier and lower-friction for AI agents to use. The implementation details are all in the repo, and everything is transparent and fully open source.
This is exactly what I've been looking for. I use browser automation heavily for B2B workflows and the local-first approach with real Chrome profiles is a game changer - no more fighting with headless detection. The MCP server integration is a nice touch too. How does tab claiming work when multiple agents need to coordinate?
Open Browser Use
@nossa_iyamu Thanks! OBU is intentionally closer to the tool/infra layer. The coordination policy is left to the upper layer, so multiple agents or callers can define how they want to claim and hand off groups/tabs, including tabs owned by another agent or a human. My suggestion is to customize the SKILL.md for your own workflow and make those coordination rules explicit there.
Good to see more OSS browser automation. The piece that's been hardest for us is fingerprint consistency across a long-running agent session — Cloudflare clocks the drift fast. How are you handling persistent identity?
Browser agents become really powerful once they can reliably navigate authentication flows and multi-tab sessions. I like the local/open-source direction here since debugging agent behavior is still a major challenge in production environments