seek://map/dsh-browser-control
dsh-browser-control
#browser #playwright #automation
TL;DR
A real, visible Chrome under the agent's fingers: per its README, ~80 browser_* tools via Playwright MCP, with a live screenshot panel in the GUI that refreshes after every action.
An agent with a visible browser — watch it click, type and scroll a real Chrome, with a live viewport pinned in your GUI.
Key points
- Drives an actual visible Chrome, not a headless shadow
- Per its README, ~80 browser_* tools: navigate, click, type, snapshot and more
- Embedded screenshot panel refreshes after each action
FAQ
Why a visible browser?
You can watch and intervene mid-task — and sites behave like real traffic.
What can it automate?
Navigation, clicking, typing, DOM snapshots and the long tail of browser_* operations.
Does it need a separate MCP server?
It wires Playwright's MCP server into DSH; setup is documented in the repository.
Official references