jev-browser-use-mcp
Related Servers
Alternatives to jev-browser-use-mcp
No user-submitted related servers found.
Related Servers
- AlicenseNot gradedqualityBmaintenanceEnables agents to perform typed, site-specific actions across a signed-in Chrome browser, interacting with services like Gmail, Reddit, Drive, and Amazon through natural language requests.1MIT
- AlicenseNot gradedqualityCmaintenanceGives coding agents full control over a real Chrome browser, enabling navigation, clicking, typing, screenshotting, and more via an MCP server and Chrome extension.16 npm2MIT
- AlicenseNot gradedqualityCmaintenanceEnables agents to control a real Chromium browser with semantic tools, providing compact observations and outcome-verified actions for web interaction tasks.1MIT
- AlicenseNot gradedqualityBmaintenanceEnables any coding agent to use the user's existing Chrome profiles and logins to open tabs, read pages, fill forms, click through flows, and take screenshots.1,430 npmMIT
- AlicenseNot gradedqualityBmaintenanceEnables coding agents to control and inspect a live Chrome browser for reliable automation, in-depth debugging, and performance analysis.1,060,392 npmApache 2.0
- AlicenseNot gradedqualityBmaintenanceEnables coding agents to retrieve verified web evidence via headless Chromium rendering, sitemap navigation, and search candidate discovery, returning clean Markdown, structured JSON, and verbatim quoted proof.3 npmMIT
TDQS
Scored across 2 tools
The two tools have clearly distinct responsibilities: one drives and continues browser sessions, the other closes them. There is no overlap or ambiguity between starting/continuing work and cleaning up a session.
Both names follow a consistent verb_noun snake_case pattern with the browser prefix: run_browser_task and close_browser_session. The naming is predictable and easy to reason about.
Two tools is on the thin end, but run_browser_task intentionally subsumes a broad range of browser actions through a goal-based interface. It is reasonable for a minimal server, though the surface feels slightly limited.
The core browser-use lifecycle is covered: start a task, continue/redirect it, and close the session. Minor gaps exist, such as no explicit session listing or status tool, but the natural-language task abstraction covers most practical workflows.