v0.9.5 — latest stable

The local MCP server that gives
any agent safe desktop control

Any model. Any app. One MCP entry. Local-only. 6 compact tools, single safety chokepoint, no telemetry. v0.9: MCP-only protocol, Reflector feedback, guides marketplace.

View on GitHub Quick Start
6
Compact Tools (public surface)
97
Granular Tools (compat / debug)
3
Operating Systems
Two ways to use it

Run it yourself, or hand it to your agent.

Test from the CLI

Plain English in, actions out.

clawdcursor doctor clawdcursor agent

Wire it into your agent

One MCP entry, desktop control appears as native tools.

Claude Code Cursor Windsurf OpenClaw Zed
Pick a mode

How will your AI talk to it?

Same tools, three entry shapes. Pick once during install.

clawdcursor mcp — recommended

AI lives in your editor (Claude Code, Cursor, Windsurf, Zed). Editor spawns clawdcursor on demand over stdio. No daemon, no port.

{
  "mcpServers": {
    "clawdcursor": {
      "command": "clawdcursor",
      "args": ["mcp", "--compact"]
    }
  }
}
6 / 97
Compact / Granular tools
stdio
Transport

clawdcursor agent — autonomous daemon

clawdcursor brings its own LLM brain (configured via doctor). For unattended runs, scheduled tasks, multi-process orchestration.

  1. Run clawdcursor doctor · pick a provider
  2. Run clawdcursor agent
  3. POST tasks to 127.0.0.1:3847/mcp
:3847
HTTP MCP
13+
Providers

clawdcursor agent --no-llm — BYO brain

Your agent already has a brain — you just want HTTP tools. Same daemon, no built-in pipeline.

  1. Run clawdcursor agent --no-llm
  2. 97 tools on :3847/mcp
  3. Stateless — no session init needed
97
Granular tools (compat)
any
HTTP client
How it works

Cheap paths first.

A11y tree before pixels. Vision only when needed.

1 Router

Zero LLM

Pattern-match shortcuts for common tasks. Sub-second.

2 Agent

One loop, three modes

Blind reads the a11y tree. Hybrid adds on-demand screenshots. Vision is the fallback for canvas UIs.

3 Safety

Single chokepoint

Every tool call gates through one safety layer. Destructive actions need confirmation.

🎯

Compact tool surface

6 compound tools — computer, accessibility, window, system, browser, task. ~12× smaller catalog than the granular surface.

🧩

One adapter per OS

Windows, macOS, Linux behind a single interface. Linux covers X11 and Wayland.

Features

Any OS. Any model.

🍎

macOS

TCC-safe. clawdcursor grant handles Accessibility + Screen Recording.

🪟

Windows

Native UIA + Windows.Media.Ocr. x64 and ARM64.

🐧

Linux

X11 and Wayland. AT-SPI for a11y, Tesseract for OCR.

🖱️

Smart tools

Click by name, type by label, read screen. A11y first, OCR as fallback.

⌨️

Shortcuts engine

Platform-aware key combos — Cmd on macOS, Ctrl elsewhere. No LLM cost.

📖

Guides marketplace

Community shortcuts & workflows for popular apps. Fetched on demand. Contribute via PR →

Guides Marketplace

Community knowledge for popular apps.

The agent reasons from scratch on unfamiliar apps. For popular ones, fetch the shortcuts.

⚡

5–10× faster on known apps

Guides ship keyboard shortcuts, workflow patterns, layout cues, and failure modes. The agent skips the discovery loop.

🌐

Fetched on demand

Cached locally for 7 days, LRU at 50 entries. Agent never blocks on the network — offline falls back to first-principles reasoning.

🛡️

Linter every load

Schema check + prompt-injection patterns + dangerous-prose detection on every guide before injection. Defense-in-depth.

10 verified guides at launch

discord · excel · figma · gmail · mspaint · olk (new Outlook) · outlook · slack · spotify · youtube (19 workflows, 36 shortcuts, 13 tips) 

# browse + manage cache
clawdcursor guides available
clawdcursor guides install youtube
clawdcursor guides list

# submit your own
clawdcursor guides lint my-app.json
clawdcursor guides submit my-app.json

Source: github.com/AmrDab/clawdcursor-guides · Served at clawdcursor.com/app-guides · PRs via GitHub · Ratings via 👍/👎 on vote: <app> issues

Tools 6 compact tools + 97 granular The 6 compact compounds are the recommended public surface. Each row lists the actions you pass via { "action": "…" }. The 97 granular tools (one schema per verb) are listed below for compatibility and debugging — use them when your runtime requires every primitive as a top-level MCP tool.
Compound Purpose Actions
computer Mouse, keyboard, screenshots. The raw I/O surface. screenshot · click · double_click · right_click · triple_click · hover · scroll · scroll_horizontal · drag · drag_path · type · key · wait
accessibility Drive UI by element name, not by pixel. Survives DPI, resize, layout shifts. read_tree · find · get_element · focused · invoke · focus · set_value · get_value · expand · collapse · toggle · select · state · list_children · wait_for
window Launch, focus, resize. App-level state management. list · active · focus · maximize · minimize · restore · close · resize · list_displays · screen_size · open_app · open_file · open_url · switch_tab · navigate
system Clipboard, OCR, shortcuts — plus pipeline introspection (classify, app-guide, system-prompt). The meta surface for an external brain. clipboard_read · clipboard_write · system_time · ocr · undo · shortcuts_list · shortcuts_run · delegate · detect_webview · relaunch_with_cdp · app_guide · detect_app · classify_task · system_prompt
browser Chrome DevTools Protocol — real DOM access for Electron / WebView2 apps whose a11y tree is sparse. connect · page_context · read_text · click · type · select_option · evaluate · wait_for · list_tabs · switch_tab · scroll
task Hand off the whole task to clawdcursor's autonomous loop. Daemon mode only — requires clawdcursor agent with an LLM configured. single arg: { instruction: string } — no action enum

Compact form (recommended): computer({ "action": "key", "combo": "mod+s" }) — ~1,500 tokens of catalog. Granular form (compat / debug): key_press({ "key": "mod+s" }) — 97 individual tools, one schema per verb. Both produce identical effects through the same safety.evaluate() chokepoint. Pass --granular (instead of --compact) to expose the granular surface over MCP. See schema.snapshot.json for every parameter.

CLI Every command The CLI is for humans diagnosing an install or managing the guide cache. Agents connect via MCP. 
# Install & setup
clawdcursor consent          # one-time desktop-control authorization
clawdcursor grant            # macOS Accessibility + Screen Recording prompts
clawdcursor doctor           # verify permissions, configure AI provider
clawdcursor status           # readiness check (consent, permissions, AI config)

# Run
clawdcursor mcp              # stdio MCP server for editor hosts
clawdcursor mcp --compact    # same, with 6 compound tools (recommended)
clawdcursor agent            # HTTP MCP daemon at :3847/mcp, optional built-in LLM
clawdcursor agent --no-llm   # tool surface only — your agent brings its own brain
clawdcursor stop             # stop every running mode
clawdcursor uninstall        # remove all clawdcursor config and data

# Guides — see the Guides Marketplace section above, or run
# `clawdcursor guides --help` for the full set.
Install

One line per OS.

Providers auto-detected.

PowerShell
powershell -c "irm https://clawdcursor.com/install.ps1 | iex"
clawdcursor consent --accept   # one-time desktop-control authorization (required)
clawdcursor doctor            # verify permissions; optionally configure an LLM
clawdcursor mcp               # OR `clawdcursor agent` — see "Pick a mode" above

Node.js 20+. Localhost only, bearer-token auth on every HTTP request. Installer clones into ~/clawdcursor, builds, and npm links a global shim. Pin a version with $env:VERSION='v0.9.5' (PowerShell) or VERSION=v0.9.5 (bash) before running the installer.

Give your agent safe desktop control.

Open source. Any model. Localhost only. No telemetry.

Star on GitHub