← Back to directory
C

Clawd Cursor

Community
Safe cross-OS desktop control for any AI agent — the fallback execution layer.
GitHub source repository ↗
★ 398 Stars Category · Other Very popular Source revision 73027f7d4032
68FMRS · C
Reliability
10/20
Security and permissions
13/20
Maintenance
15/20
Documentation
16/20
Setup experience
14/20

Clawd Cursor is a powerful MCP server that provides safe desktop automation via accessibility tree and OCR, with verification and a single safety gate. It's suitable for agents needing to control native apps, but requires proper permissions and may have a learning curve.

Read the FMRS scoring method →

Clawd Cursor is a local MCP server that gives any tool-calling agent (Claude Code, Cursor, Windsurf, OpenClaw, etc.) safe control of the real desktop. It compiles the screen into a UI map by fusing the accessibility tree and OCR, tagging each element with a stable el_NN id, and acts on elements by id rather than pixel coordinates. It verifies its own actions and routes every call through a single safety checkpoint. Works on Windows, macOS, and Linux, fully local with no cloud dependency.

Tools

computer
Performs mouse and keyboard actions like clicking, typing, scrolling, dragging, and waiting.
accessibility
Reads accessibility tree, finds elements, invokes actions, sets values, compiles UI map.
window
Manages windows and apps: list, focus, resize, open app, open file/URL, switch tabs.
system
System-level operations: clipboard, OCR, undo, detect webview, run shortcuts.
browser
Controls browser via CDP: connect, read text, click, type, evaluate, manage tabs.
task
Delegates a task to the built-in thin agent loop.
batch
Collapses multiple calls into one round-trip, each step still safety-gated.
compile_ui
Fuses accessibility tree and OCR into a confidence-scored map of UI elements.
find_button
Finds a button by name or text and returns its element ID.
find_field
Finds an input field by name or label and returns its element ID.
smart_click
Smart click: tries accessibility tree first, falls back to OCR or vision if needed.
smart_type
Smart type: tries accessibility tree first, falls back to OCR or vision if needed.
smart_read
Smart read: tries accessibility tree first, falls back to OCR or vision if needed.
detect_webview
Detects Electron/WebView2 apps and switches to browser control.

Setup

  1. Install globally: npm i -g clawdcursor
  2. Grant consent: clawdcursor consent --accept
  3. macOS only: grant permissions with clawdcursor grant
  4. Add to MCP config:

{"mcpServers":{"clawdcursor":{"command":"clawdcursor","args":["mcp","--compact"]}}}

claude_desktop_config.json
{
  "mcpServers": {
    "clawdcursor": {
      "command": "clawdcursor",
      "args": [
        "mcp",
        "--compact"
      ]
    }
  }
}

Fit and risk

Best for

  • AI agents needing cross-platform desktop automation
  • Scenarios requiring safety controls with allow/confirm/block
  • Privacy-sensitive tasks that should stay local
  • Agents wanting to avoid high vision model costs

Not for

  • Browser-only automation where specialized tools exist
  • Cloud-hosted or non-local execution
  • Tasks requiring strict human approval
  • Pure pixel-based manipulation without vision models

Required permissions

  • Desktop control (click, type)
  • Read screen content (accessibility tree, OCR, screenshots)
  • Open apps and files
  • Browser control (CDP)
  • Clipboard read/write
  • System-level hotkeys

Risks and side effects

  • Agent may perform unintended actions, leading to data loss or system changes
  • Sensitive information on screen could be read by the agent
  • Potential for malicious actions if agent misbehaves
  • Requires proper permission grants, otherwise may not work
  • Confirmation mode may interrupt automation flow

Troubleshooting

  1. Ensure Node.js 20+ is installed
  2. Ensure desktop access permissions have been granted (macOS Accessibility)
  3. Check OS dependencies: Linux requires tesseract-ocr, python3-gi, etc.
  4. Verify the 'clawdcursor' command is in PATH
  5. Check logs: CLAWD_LOG=pretty for detailed operations
  6. If agent can't connect, ensure HTTP mode token is correct

Use cases

Driving native apps (e.g., Outlook, Notepad) for complex GUI tasks
Automating browser actions via CDP for Chrome/Edge
Interacting with legacy applications that have no API or CLI
Canvas-based apps when spatial reasoning is required
As a last-resort execution layer after native APIs and CLIs

Supported clients

Claude DesktopFull support
Claude CodeFull support
CursorFull support
WindsurfFull support
ZedFull support
OpenAI CodexFull support