$ loading_
Let AI see and control a Linux desktop for visual task automation.
This is an open-source MCP server focused on local Linux Wayland desktop automation, including screenshots, OCR, and mouse/keyboard control, with no declared secrets or remote endpoints. Its main concerns are local control and data exposure on the host rather than explicit exfiltration, so the overall posture is caution.
The material explicitly states that no keys or environment variables are required, and no API tokens, account credentials, or cloud authentication are mentioned, so direct credential exposure appears low.
Neither the material nor the objective checks declare any remote endpoints; the description focuses on local Wayland desktop screenshots, OCR, icon detection, and input control, with no concrete indication of sending user data to external services.
The system marks it as executes-code, and its mouse/keyboard control capability indicates it can perform local actions and drive desktop interaction on the host. This is a high-privilege local automation characteristic typical of such MCP tools and warrants containment and scope limits.
Its stated screenshot, OCR, and desktop visual parsing features imply access to on-screen content and possible indirect interaction with applications and data in the user session. The material does not specify file read/write scope, and there is no clear evidence of permissions beyond its stated function, but screen contents themselves may be sensitive.
A positive factor is that there is an open-source repository and the source is in principle auditable; however, it comes from a third-party registry, has no declared license, 0 stars, unknown maintenance status, and no README, leaving limited evidence of maturity and trustworthiness and making code/dependency review advisable.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "screen-mcp" yet — see the docs or source repo.
Connect screen-mcp to my Wayland desktop, open the target app, click through login and the settings page, capture screenshots at each step, check for any error dialogs, and output a test summary.
A UI test report with action steps, key screenshots, detected issues, and a final test result.
Use screen-mcp to detect the browser and spreadsheet app icons on screen, open them, copy specified data from a webpage, and paste it into the matching spreadsheet columns; if the UI layout changes, first use screenshots, OCR, and icon detection to locate elements.
A completed cross-app automation run with execution logs, detected UI elements, and a result summary.
Capture the current Linux desktop, identify all visible buttons, input fields, and icons, extract on-screen text, and explain where the user should click next to complete a file upload.
A visual analysis with OCR text, UI element locations, and recommended next-step actions.
Capture screen frames on demand for multimodal Q&A and interface understanding.
Capture screens, extract text, and automate clicks on macOS interfaces.
Enable AI to automate macOS desktop actions like screenshot, click, typing, and scrolling.
Enable AI agents to fully control Linux desktops and manage system tasks.
Let AI capture Windows screenshots for troubleshooting, review, and visual context.
Let AI agents discover macOS windows and capture screenshots for UI debugging.