About

Agents are asked to work on the web every day, and most browser tooling still hands them a DOM and hopes for the best. BrowserControl exists to close that gap: show the agent the page, let it point at what it wants, and keep the whole loop local.
Every design decision is measured against two questions. Does it make an agent more capable in a real browser? And does it stay out of the user's way?
How it's built
| Language | Python 3.11+ |
| Built on | FastMCP, Playwright, Pillow |
| Browser | Chromium, or your own Chrome over CDP |
| Quality | mypy strict, ruff, bandit; tests on Linux, Windows, and macOS |
| License | MIT |
Links
- Source: github.com/adityasasidhar/browsercontrol
- Package: pypi.org/project/browsercontrol
- Issues: GitHub issues
- Release notes: Changelog
Maintainer
BrowserControl was created and is maintained by Aditya Sasidhar. Contributions are welcome; see Contributing.
