Screen Control

Gives AI agents safe computer use on Windows: screen perception via OCR, single frames and a live MJPEG stream, plus safety-gated mouse, keyboard, window, virtual-desktop and game-mode control over a local token-authenticated HTTP API and MCP (16 tools).

GitHub

Dùng thử MCP nàyĐược tài trợ

Tài liệu

🖥️ Screen Control

A local remote-control system for AI agents: watch your computer's screen live and send mouse/keyboard commands to it. Everything runs on your own machine — no data ever leaves it, no cloud middleman.

An AI agent typing into Notepad through the screen-control API

_{Every keystroke and every screenshot in this GIF went through the API —
the agent never touched a physical keyboard.}

Why Screen Control?

AI agents today can write code and call APIs — but they can't see or touch your desktop. Screen Control gives any agent general-purpose computer use over a clean, safety-gated HTTP/MCP interface:

Perceive — OCR for text, single frames or a live MJPEG stream for vision-capable models, and a text-only diff endpoint for models that can't consume images at all.
Act — absolute and relative mouse, Unicode-safe keyboard, window management, background (focus-free) control, virtual desktops.
Stay safe — token auth, blocked deadly shortcuts, focus guard, a stuck-input watchdog and an emergency failsafe are all enforced server-side, no matter how confused the agent gets.

One process, zero configuration, works with any language that can speak HTTP — or natively through MCP in Claude Desktop, Cursor, VS Code and cloud agents.

Performance Is Agent-Bound

Screen Control is the perception and actuation layer — the eyes and hands. The effective speed and capability of any agent using it are bounded by that agent itself and by the environment it runs in:

Thinking speed — one action per agent "turn": the perceive → plan → act → verify loop lives in the agent, so model inference latency and reasoning depth directly set the pace. The API itself adds only milliseconds per call.
Context capacity — screen readings (OCR text, frames, diffs) consume the agent's context window; a larger window means more situational awareness before verification degrades.
Runtime environment — network latency, MCP/HTTP round-trip overhead, tool-call limits and hosting constraints all stack on top of the loop.

In practice this means: the same repo makes a fast reasoning model fast and capable, and makes a slow model slow — the toolchain is not the bottleneck. Real-time or action-heavy tasks need an agent with fast inference and tight tool-loop latency; slower agents should prefer deliberate, verification-heavy tasks.

Why Screen Control?
Features
Architecture
Installation
Quick Start
API Reference
🤖 For AI Agents
MCP Support (One-Click Cloud Agents)
Security Model
Game Mode Guide
Vision Access Guide
Troubleshooting
Testing
Contributing
License

Features

Feature	Description
🖼️ Live screen feed	Continuously refreshing screenshot in the browser
🖱️ Mouse control	Click, right-click, double-click, scroll, drag & drop via live screenshot
⌨️ Keyboard control	Text typing (Unicode/Turkish included, layout-independent), keys and shortcuts (Ctrl+C, Alt+Tab…)
👁️ OCR	Converts on-screen text to machine-readable format
📷 Vision access	Raw-pixel paths for image-capable models: single frames, MJPEG stream, text-based motion detection
🪟 Window management	List, focus, safe close (WM_CLOSE), kill (task-manager style)
🖥️ Focus-free control	Read/write background windows via PostMessage without stealing focus
🎮 Game mode	Camera look via relative mouse movement, hold-to-move keys
🔐 Token auth	Every request requires `X-Auth-Token` (CSRF protection)
🦺 Stuck-input watchdog	Auto-releases held keys after 30 s of inactivity
🛟 Failsafe	Cursor to top-left corner aborts all commands (disabled in game mode)

Architecture

┌─────────────────────────────────────────────────────────┐
│                    Browser (Web UI)                      │
│  ┌──────────┐  ┌──────────┐  ┌────────────────────────┐ │
│  │  Live     │  │  Control  │  │  Windows / Game Mode   │ │
│  │  View     │  │  Panel    │  │  Panel                 │ │
│  └────┬─────┘  └────┬─────┘  └───────────┬────────────┘ │
│       │              │                     │              │
└───────┼──────────────┼─────────────────────┼──────────────┘
        │              │                     │
        ▼              ▼                     ▼
┌─────────────────────────────────────────────────────────┐
│                  HTTP API (Flask)                        │
│                  127.0.0.1:8745                           │
│                                                         │
│  /api/screenshot    /api/mouse     /api/key              │
│  /api/vision/*      /api/ocr       /api/window           │
│  /api/game          /api/held      /api/release_all      │
│  /api/windows       /api/desktops  /api/desktop          │
│                                                         │
│  ┌─────────────┐  ┌──────────────┐  ┌──────────────┐   │
│  │ Auth Layer  │  │  Watchdog    │  │  OCR Engine  │   │
│  │ (token)     │  │  (30s auto)  │  │  (RapidOCR)  │   │
│  └─────────────┘  └──────────────┘  └──────────────┘   │
└─────────────────────────────────────────────────────────┘
        │              │                     │
        ▼              ▼                     ▼
┌─────────────────────────────────────────────────────────┐
│                  control.py (Core)                        │
│                                                         │
│  Screen:  mss (fast capture), PIL (processing)          │
│  Mouse:   pyautogui (absolute), SendInput (relative)    │
│  Keyboard: pyautogui + SendInput+KEYEVENTF_UNICODE      │
│  Windows: Win32 API (EnumWindows, SetForegroundWindow)   │
│  Background: PrintWindow (capture), PostMessage (input)  │
│  Virtual Desktops: pyvda                                 │
│  Game Mode: ClipCursor + MOUSE_MOVE_RELATIVE             │
└─────────────────────────────────────────────────────────┘
┌──────────────────────────────────────────────┐
│              backends/ (pluggable)           │
│  WindowsBackend  │ LinuxBackend │ MacOSBackend│
│      (full)      │   (stub)     │   (stub)   │
└──────────────────────────────────────────────┘

Coordinates & Concurrency

Per-Monitor DPI awareness. control.py calls SetProcessDpiAwarenessContext(PER_MONITOR_AWARE_V2) at import time — before the pyautogui import, because pyautogui touches coordinate APIs during import and would otherwise lock the process to the interpreter manifest's default (system-aware). With PMv2 active, every coordinate in the system is a physical pixel end to end: mss capture, OCR bounding boxes, pyautogui/SendInput clicks, ClipCursor. On High-DPI displays (125%/150% scaling) nothing drifts between what OCR reports and where the mouse clicks.

Lock architecture. The server uses two independent locks instead of one global lock:

Lock	Protects	Endpoints
`_input_lock`	mouse, keyboard, game mode, window ops	`/api/mouse`, `/api/key`, `/api/game`, `/api/window/post`, ...
`_read_lock`	capture, OCR, vision, enumeration	`/api/screenshot`, `/api/ocr`, `/api/vision/*`, `/api/windows`, ...

A slow OCR (3–5 s on a busy screen) no longer freezes concurrent screenshot or vision reads — reads queue behind reads, inputs behind inputs.

Live-Loop Working Principle

This system is designed for a live perceive-act loop, not pre-written command chains:

READ — OCR or vision reads the screen before and after every action
ONE ACTION — each round sends a single command
VERIFY — acceptance is "it appeared on screen", not "I sent it"
ADAPT — if verification fails, the next step changes based on what is actually seen

This is enforced by the expect_hwnd guard: typing is refused (409) if the foreground window doesn't match the target.

Installation

cd screen-control
pip install -r requirements.txt

Requirements

Package	Purpose	Required?
`mss`	Fast screen capture	✅ Yes
`pyautogui`	Mouse/keyboard control	✅ Yes
`pyvda`	Virtual desktop management	✅ Yes
`flask`	HTTP server	✅ Yes
`Pillow`	Image processing	✅ Yes
`rapidocr-onnxruntime`	OCR (screen text reading)	⚠️ Optional

Note: The OCR package is large and may take a while to install. If it fails, everything else still works — only the OCR feature is unavailable.

System Requirements

OS: Windows 10/11 (x64)
Python: 3.10+
Display: Any resolution; the system adapts automatically

Quick Start

# 1. Start the server
cd screen-control
python server.py

# 2. Open in browser
#    http://127.0.0.1:8745

# 3. Or control via API
TOKEN=$(cat .token)
curl -H "X-Auth-Token: $TOKEN" http://127.0.0.1:8745/api/screenshot -o screen.jpg

API Reference

Capabilities

Returns the active backend name and what it can do. Agents should call this first (see ROADMAP.md for the multi-platform plan).

GET /api/capabilities
→ {"ok": true, "backend": "windows",
   "capabilities": {"screen_capture": true, "game_mode": true, ...}}

Feature values: true (supported), false (absent), null (unknown — stub backend), "optional" (depends on an optional dependency).

Platform Support Matrix

Capability	Windows	Linux X11	Linux Wayland	macOS
Screen capture	Full	Full	Portal-dependent	Permission required
OCR	Full/optional	Full/optional	Full/optional	Full/optional
Mouse control	Full	Full	Restricted	Accessibility permission
Keyboard control	Full	Full	Restricted	Accessibility permission
Window enumeration	Full	WM-dependent	Limited	Accessibility/API-dependent
Background input	Strong	WM/app-dependent	Usually unavailable	Limited
Virtual desktops	Supported	DE/WM-dependent	DE/WM-dependent	Spaces-specific
Game mode	Supported	Experimental	Limited	Experimental

Linux and macOS backends are currently fail-closed stubs: every operation returns BACKEND_UNAVAILABLE (501) until implemented (ROADMAP Phases 5–7). Windows is the reference backend.

Authentication

Every request must include the X-Auth-Token header. The token is generated on each server start and written to .token.

TOKEN=$(cat .token)

Code	Meaning
401	Missing or invalid token
415	POST without `Content-Type: application/json`

Token bootstrap (for the bundled web UI):

GET /token
→ {"ok": true, "token": "abc123..."}

The /token endpoint is safe: Same-Origin Policy prevents foreign pages from reading it.

Screen Capture

`GET /api/screenshot`

Returns a JPEG screenshot.

Parameter	Type	Default	Description
`monitor`	int	1	Monitor index
`region`	string	—	`x,y,w,h` sub-region

curl -H "X-Auth-Token: $TOKEN" -o screen.jpg http://127.0.0.1:8745/api/screenshot
curl -H "X-Auth-Token: $TOKEN" "http://127.0.0.1:8745/api/screenshot?region=0,0,800,600"

`GET /api/info`

Returns screen dimensions and system state.

{"ok": true, "width": 1920, "height": 1080, "ocr_available": true,
 "failsafe": true, "game_mode": false}

Mouse Control

`POST /api/mouse`

action	Required params	Optional params	Description
`move`	`x`, `y`	`duration` (default 0.15)	Move cursor to absolute position
`click`	`x`, `y`	`button` (left/right), `clicks` (default 1)	Click at position
`scroll`	`clicks`	`x`, `y`	Scroll wheel (positive=up)
`drag`	`x1`, `y1`, `x`, `y`	`duration`, `button`	Drag between two points
`down`	`button` (default "left")	—	Press and hold mouse button
`up`	`button` (default "left")	—	Release held mouse button

# Click at center of screen
curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"click","x":960,"y":540,"button":"left"}'

# Right-click
curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"click","button":"right","x":960,"y":540}'

# Scroll down
curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"scroll","clicks":-3}'

Keyboard Control

`POST /api/key`

action	Required params	Description
`press`	`key`	Press and release a key
`down`	`key`	Hold a key down (tracked for watchdog)
`up`	`key`	Release a held key
`hotkey`	`keys` (array)	Key combination (e.g. `["ctrl","c"]`)
`type`	`text`	Type text (Unicode, layout-independent)

Optional param	Default	Description
`expect_hwnd`	—	Window handle to verify focus (409 if mismatch)
`interval`	0.03	Delay between characters for `type`

# Press Enter
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"press","key":"enter"}'

# Ctrl+C
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"hotkey","keys":["ctrl","c"]}'

# Type text (Turkish characters supported)
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"type","text":"Merhaba dünya"}'

# Hold W key down (for walking in games)
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"down","key":"w"}'

OCR (Screen Reading)

`POST /api/ocr`

Converts on-screen text to machine-readable format.

Param	Type	Default	Description
`region`	array	—	`[x, y, w, h]` sub-region (faster)

{
  "ok": true,
  "text": "Hello World\nFile Edit View",
  "lines": ["Hello World", "File Edit View"],
  "items": [
    {"text": "Hello World", "x": 960, "y": 40},
    {"text": "File Edit View", "x": 100, "y": 15}
  ]
}

# Full screen OCR
curl -X POST http://127.0.0.1:8745/api/ocr -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{}'

# Region-only (faster, ~10x for small regions)
curl -X POST http://127.0.0.1:8745/api/ocr -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"region":[0,0,800,100]}'

Vision Access (Image Models)

Three endpoints for models that can consume images:

Endpoint	Description
`GET /api/vision/frame`	Single JPEG frame (raw or base64)
`GET /api/stream`	MJPEG live stream
`POST /api/vision/diff`	Text-based motion detection (no vision needed)

`GET /api/vision/frame`

Param	Default	Description
`scale`	1.0	Downscale factor (0.5 = half size)
`gray`	0	1 for greyscale
`quality`	80	JPEG quality (20-95)
`format`	—	`base64` for JSON response
`region`	—	`x,y,w,h` sub-region

# Half-size greyscale frame as base64 (for text-only models)
curl "http://127.0.0.1:8745/api/vision/frame?scale=0.5&gray=1&format=base64" \
  -H "X-Auth-Token: $TOKEN"

`GET /api/stream`

MJPEG live stream. Drop into <img src> or consume frame-by-frame.

Param	Default	Description
`fps`	10	Frames per second (1-30)
`quality`	70	JPEG quality
`scale`	1.0	Downscale factor
`region`	—	`x,y,w,h` sub-region

`POST /api/vision/diff`

Text-based motion detection — no vision model required.

Body	Description
`{}`	Compare against last stored frame
`{"grab":"gray"}`	Store current frame for next comparison
`{"b64_prev":"..."}`	Compare against provided previous frame

{
  "ok": true,
  "changed": true,
  "changed_pct": 12.5,
  "bbox": [100, 200, 400, 350],
  "tiles": [
    {"row": 2, "col": 4, "pct": 35.2, "center": [1000, 390]}
  ]
}

Window Management

`GET /api/windows`

List all visible windows.

{
  "ok": true,
  "windows": [
    {
      "hwnd": 123456,
      "title": "My Application",
      "process": "app.exe",
      "pid": 7890,
      "focused": true,
      "rect": [0, 0, 1920, 1080],
      "desktop": 1
    }
  ]
}

`POST /api/window`

action	Required	Optional	Description
`focus`	`hwnd`	—	Bring window to foreground
`close`	`hwnd`	`expect_title`, `expect_process`	Safe close via WM_CLOSE
`kill`	`hwnd`, `pid`	—	Force kill (task-manager style)
`topmost`	`hwnd`	—	Set always-on-top
`untopmost`	`hwnd`	—	Remove always-on-top
`maximize`	`hwnd`	—	Maximise window

# Focus a window
curl -X POST http://127.0.0.1:8745/api/window -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"hwnd":12345,"action":"focus"}'

# Safe close (with title verification)
curl -X POST http://127.0.0.1:8745/api/window -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"close","hwnd":12345,"expect_title":"Notepad"}'

# Kill process
curl -X POST http://127.0.0.1:8745/api/window -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"kill","hwnd":12345,"pid":7890}'

Focus-Free (Background) Control

Read and control windows without stealing focus — the user keeps working on their main desktop.

`GET /api/window/capture`

Capture a window via PrintWindow (works even on another virtual desktop).

Param	Description
`hwnd` (required)	Window handle
`client`	1 = client area only
`ocr`	1 = return OCR text instead of image

# Capture window as PNG
curl "http://127.0.0.1:8745/api/window/capture?hwnd=12345" \
  -H "X-Auth-Token: $TOKEN" -o window.png

# Capture + OCR in one call
curl "http://127.0.0.1:8745/api/window/capture?hwnd=12345&ocr=1" \
  -H "X-Auth-Token: $TOKEN"

`POST /api/window/post`

Send input to a window, choosing the delivery path automatically.

action	Description
`type`	Type text (Unicode-safe)
`key`	Send a key press
`hotkey`	Send a key combination
`click`	Click at client coordinates
`scroll`	Scroll the window
`drag`	Drag inside the window

Optional mode parameter controls routing:

mode	Behavior
`auto` (default)	Decided by input-mode probe (see below)
`background`	Force PostMessage path (window keeps focus/z-order)
`focused`	Force focus + SendInput path

# Type into a background Notepad
curl -X POST http://127.0.0.1:8745/api/window/post -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"hwnd":12345,"action":"type","text":"Hello from background!"}'

Routing rules (mode=auto):

postmessage — classic Win32 app: background PostMessage, no focus change.
uia — WinUI/UWP/XAML surface (single DirectX canvas, no Win32 child controls): posted messages are silently swallowed, so the window is focused and the action is replayed through SendInput (client coords converted to screen). This is the documented fallback for modern apps.
focused — window is already foreground: focused SendInput path.
invalid — HTTP 409; not a reachable top-level window.

`GET /api/window/input-mode`

Classify how a window receives input before posting to it. Returns one of focused | postmessage | uia | invalid.

curl "http://127.0.0.1:8745/api/window/input-mode?hwnd=12345" \
  -H "X-Auth-Token: $TOKEN"

WinUI note: New Notepad (and other XAML-hosted apps) has no classic child Edit control to post to — the whole UI is one DirectX surface. input-mode reports uia for these; /api/window/post then automatically uses the focused SendInput path. /api/window/children remains useful for classic apps with real child controls.

`GET /api/window/children`

List child controls of a window (class name + title + hwnd).

curl "http://127.0.0.1:8745/api/window/children?hwnd=12345" -H "X-Auth-Token: $TOKEN"

Virtual Desktops

`GET /api/desktops`

List all virtual desktops.

`POST /api/desktop`

action	Params	Description
`switch`	`number`	Switch to desktop N
`create`	—	Create a new desktop

curl http://127.0.0.1:8745/api/desktops -H "X-Auth-Token: $TOKEN"

curl -X POST http://127.0.0.1:8745/api/desktop -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"switch","number":2}'

Game Mode

action	Params	Description
`start`	`sensitivity` (default 12)	Lock cursor to center, enable game input
`move`	`dx`, `dy`, `sensitivity`	Rotate camera (relative mouse)
`stop`	—	Release cursor + all held input
`heartbeat`	—	Keep-alive for long holds

# Start game mode
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"start","sensitivity":12}'

# Look right
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"move","dx":50,"dy":0}'

# Hold W to walk forward
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"down","key":"w"}'

# ... later ...
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"up","key":"w"}'

# Stop game mode
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"stop"}'

Safety Endpoints

`GET /api/held`

Returns currently held keys/buttons and watchdog status.

{
  "ok": true,
  "keys": ["w", "shift"],
  "buttons": ["left"],
  "game_mode": true,
  "idle_seconds": 5.2,
  "watchdog_count": 0,
  "last_watchdog": null
}

`POST /api/release_all`

Emergency: release everything (held keys, mouse buttons, game-mode cursor lock).

curl -X POST http://127.0.0.1:8745/api/release_all -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{}'

🤖 For AI Agents

A dedicated, comprehensive guide for AI agents (LLMs, vision models, automation frameworks) is available in AGENT_GUIDE.md.

It covers:

Perceive-act loop (read → plan → act → verify)
Focus guard (expect_hwnd) to prevent wrong-window accidents
App automation and game control workflows
Vision access for image-capable models
Text-based motion detection
Bandwidth optimization
Complete curl examples

MCP Support (One-Click Cloud Agents)

Model Context Protocol (MCP) turns this project into a plug-and-play toolbox for any MCP-capable agent: Claude Desktop, Claude Code, Cursor, VS Code Copilot Agent mode, custom cloud agents — no custom glue code, no curl scripts. The agent discovers and calls the tools natively.

How it works

MCP agent (cloud or desktop)
        │  MCP protocol (stdio or streamable-HTTP)
        ▼
  mcp_server.py   ← thin wrapper: tools → HTTP calls, token auto-read
        │  REST + X-Auth-Token (localhost only)
        ▼
  server.py       ← the single source of truth:
                    auth, locks, watchdog, focus guard, all safety rules

mcp_server.py adds no new powers — every safety mechanism (auth token, input/read locks, watchdog, Alt+F4 block, focus guard, failsafe) stays enforced by server.py.

Setup

pip install mcp            # optional dependency (see requirements.txt)
python server.py           # start the REST server first (it writes .token)

The MCP server auto-reads the token from .token (or the SCREEN_CONTROL_TOKEN env var) — zero configuration.

Desktop agents (stdio transport)

Claude Desktop — claude_desktop_config.json:

{
  "mcpServers": {
    "screen-control": {
      "command": "python",
      "args": ["C:/path/to/screen-control/mcp_server.py"]
    }
  }
}

Claude Code: claude mcp add screen-control -- python C:/path/to/screen-control/mcp_server.py

Cursor / VS Code: add the same entry to their MCP config files.

Remote / cloud agents (streamable-HTTP transport)

python mcp_server.py --http --port 8751
# MCP endpoint: http://127.0.0.1:8751/mcp

The HTTP transport is token-protected: every request must carry the X-Auth-Token header (same token as the REST server). Query-string tokens (?token=...) are rejected by design — URLs leak into proxy/tunnel logs, browser history and shared links, and this token grants full desktop control. Clients that cannot send custom headers should run a local stdio mcp_server.py instead. Only GET /health is open, for liveness probes. DNS-rebinding protection is disabled on this transport deliberately — tunneled requests arrive with a foreign Host header, and the rebinding threat is already covered by the token guard.

For a cloud agent, expose it through a tunnel:

cloudflared tunnel --url http://127.0.0.1:8751
# → prints a https://<random>.trycloudflare.com URL

Then configure the agent's MCP connection with <tunnel-url>/mcp plus the token from .token as a header (X-Auth-Token).

Headerless connectors (scoped keys)

For clients that cannot send custom headers (e.g. web connectors that only take an endpoint URL), create a scoped API key — a persistent, optionally time-limited credential — and embed it in the URL path:

# Create a 24-hour scoped key (requires the server to be running)
curl -X POST http://127.0.0.1:8745/api/keys -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"create","name":"spark","expires_in_hours":24}'

# Connector endpoint becomes:
#   https://<tunnel-url>/mcp/<scoped-key>

Design guarantees (SC-06):

The master session token is refused in URLs (403) — only scoped keys may travel there
Scoped keys expire automatically; expired keys authenticate nothing
Scoped keys are revocable by name at any moment via POST /api/keys ({"action":"revoke","name":"spark"}) — revocation takes effect immediately on every endpoint
The .apikeys file stores only SHA-256 hashes, never raw keys

⚠️ A tunnel exposes PC control to the internet. Keep the token secret, prefer short-lived tunnels and scoped keys for headerless connectors, and stop the server when not in use.

One-Command Startup (launcher + auto-tunnel)

start-server.bat automates the whole cloud setup and prints everything your cloud agent needs, ready to paste:

Downloads cloudflared.exe if missing (portable, no admin required)
Stops leftover instances from a previous run
Starts the REST server (port 8745) and the MCP HTTP server (port 8751)
Waits until both are healthy (/token and /health probes)
Starts a cloudflared quick tunnel, extracts its public URL from tunnel.log, and prints the summary:

 ============================================================
  ALL SYSTEMS RUNNING
 ============================================================
  Local REST API  : http://127.0.0.1:8745
  Local MCP       : http://127.0.0.1:8751/mcp
  Public MCP URL  : https://<random>.trycloudflare.com/mcp

  ------------------------------------------------------------
  PASTE INTO YOUR CLOUD AGENT  (MCP connector settings)
  ------------------------------------------------------------
  Endpoint : https://<random>.trycloudflare.com/mcp
  Header   : X-Auth-Token: <token>
  URL form : https://<random>.trycloudflare.com/mcp?token=<token>
             (only if the connector cannot send headers)
  ------------------------------------------------------------

stop-server.bat stops all three (REST, MCP, tunnel) in one go.

Available tools (16)

Category	Tools
Perception	`get_info`, `ocr_screen`, `screenshot` (real image block for vision models), `motion_diff`
Mouse / keyboard	`mouse`, `keyboard` (with `expect_hwnd`), `get_held`, `release_all`
Windows	`list_windows`, `focus_window`, `window_children`, `window_input_mode`, `window_post`, `window_capture_ocr`, `close_window`
Game mode	`game` (start / move / stop / heartbeat)

Which transport for whom

Consumer	Transport	Command
Claude Desktop / Cursor / VS Code (local)	stdio	`python mcp_server.py`
Claude Code	stdio	`claude mcp add ...` (above)
Cloud / remote agents	streamable-HTTP	`start-server.bat` (recommended) or `python mcp_server.py --http --port 8751` + `cloudflared tunnel --url http://127.0.0.1:8751`

Note: This project targets MCP Python SDK 2.x (MCPServer API). With SDK 1.x, replace the import with from mcp.server.fastmcp import FastMCP, Image and MCPServer with FastMCP.

Security Model

Threat: Malicious Web Pages (CSRF)

Even bound to 127.0.0.1, a malicious page in the browser can trigger non-preflighted requests (text/plain fetch, HTML form POST) to localhost. The browser blocks the response but not the request — the server would still execute the command.

Mitigation: Every request requires X-Auth-Token. A foreign page cannot read this token (Same-Origin Policy), so it cannot authenticate.

Additional layers:

POST requests must use Content-Type: application/json (415 otherwise)
This blocks form-encoded and text-plain POSTs even if the token leaked
Host header trust (DNS rebinding): when bound to loopback, requests carrying a non-loopback Host header are refused with 421 — a rebinding page that resolves its domain to 127.0.0.1 cannot read /token or call the API
/token and / responses carry Cache-Control: no-store so the credential is never persisted by browsers or proxies

Threat: Stuck Keys / Game Mode Lock

In game mode, ClipCursor pins the cursor to a 2×2 box — the classic pyautogui failsafe (cursor to top-left) does not work.

Mitigations:

Physical Esc / Alt+Tab — real hardware input; this API cannot block it, and it always works
POST /api/release_all — instant release of everything
Watchdog (automatic) — 30 s of server-side inactivity with held input triggers automatic release

Threat: Wrong Window Typing

Mitigations:

expect_hwnd guard on /api/key — if the foreground window doesn't match, typing is refused with 409
The focused path of /api/window/post verifies the focus after the focus switch and before any synthetic input (409 on mismatch) — input is never replayed into whatever window happens to be foreground
focus_window() raises on failure instead of silently returning

Threat: Dangerous Key Combos

Mitigation: Blocked at the API level (403) on every delivery path — the direct /api/key route, the background /api/window/post route (PostMessage), and the focused fallback route share one safety policy (control._assert_allowed):

Alt+F4 — the only banned Alt combo (Alt+Tab, Alt+menu are legitimate)
Win key — prevents Start menu, task switching
Ctrl+Alt+Del — system security screen
Shift+Delete style — prevents permanent deletion

Threat: Killing System Processes

Mitigations:

The process name is resolved from the PID directly (Win32 toolhelp snapshot), not from the visible-window inventory — windowless/background system processes get the same protection as visible ones
Critical system processes are blacklisted (default-deny for unknown PIDs): winlogon.exe, csrss.exe, smss.exe, services.exe, lsass.exe, svchost.exe, system, registry, dwm.exe
Optional expect_process confirmation: a mismatch aborts the kill with 409 — protects against killing a newly-reused PID

Threat: Resource Exhaustion (rogue agent / DoS)

A token-holding but misbehaving client should not be able to exhaust memory or starve the input lock.

Mitigations:

MAX_CONTENT_LENGTH = 1 MB — oversized request bodies are rejected (413)
region width/height/area and scale are bounded (400 otherwise)
text payloads are capped at 10,000 characters per input call
MJPEG streams are capped at 10 concurrent clients (429 beyond that)
Pillow decompression-bomb limit is set for client-supplied images

Network Access

The server binds to 127.0.0.1 by default. To expose it to the network:

python server.py --host 0.0.0.0  # ⚠️ anyone on the network can control this machine

Game Mode Guide

Setup

# 1. Focus the game window
curl -X POST http://127.0.0.1:8745/api/window -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"hwnd":GAME_HWND,"action":"focus"}'

# 2. Start game mode
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"start","sensitivity":12}'

Camera Look

# Look right
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"move","dx":50,"dy":0}'

# Look down
curl -X POST http://127.0.0.1:8745/api/game -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"move","dx":0,"dy":30}'

Movement

# Walk forward (hold W)
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"down","key":"w"}'

# ... walk for a while ...

# Release W
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"up","key":"w"}'

Minecraft-Specific

# Place block (right-click)
curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"action":"click","button":"right","x":960,"y":540}'

# Break block (hold left-click)
curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"down","button":"left"}'

# ... after breaking ...

curl -X POST http://127.0.0.1:8745/api/mouse -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"up","button":"left"}'

# Select hotbar slot
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"press","key":"1"}'

# Open inventory
curl -X POST http://127.0.0.1:8745/api/key -H "X-Auth-Token: $TOKEN" \
  -H "Content-Type: application/json" -d '{"action":"press","key":"e"}'

Suitability

Game Type	Suitable?	Notes
Minecraft (building)	✅ Yes	Place blocks, walk, mine
Minecraft (PvP)	❌ No	Too slow for fast combat
Turn-based games	✅ Yes	Ample time for read→act→verify
RPG / adventure	✅ Yes	Inventory, dialogue, exploration
Fast FPS	❌ No	Reaction time insufficient
Puzzle games	✅ Yes	Click-based, read-heavy

Vision Access Guide

For Image-Capable Models

If the consuming model can process images, use the vision endpoints directly:

GET /api/vision/frame?scale=0.5&gray=1&quality=70

This returns a single JPEG that the model can analyze for:

Game HUD elements (health, mana, inventory)
On-screen text (menus, chat, tooltips)
Visual scene understanding (blocks, entities, terrain)

For Text-Only Models

Use the diff endpoint for motion detection without vision:

POST /api/vision/diff {"grab":"gray"}   → first call: stores frame
POST /api/vision/diff                    → subsequent calls: returns diff

The response tells you where things changed (tile coordinates) and how much (percentage), which is sufficient for:

Detecting that an action had an effect
Locating moving elements on screen
Tracking animation state changes

Bandwidth Optimization

Approach	Payload	Use Case
`scale=1.0, gray=0`	~500 KB	Full detail
`scale=0.5, gray=1`	~50 KB	Good for most vision models
`scale=0.25, gray=1`	~10 KB	Maximum compression
`diff` (text)	~1 KB	Text-only agents
`region=...`	Variable	Focus on specific area

Troubleshooting

"OCR engine not installed"

pip install rapidocr-onnxruntime

Server won't start (port in use)

# Find the process using port 8745
netstat -ano | findstr ":8745"

# Kill it
taskkill /PID <pid> /F

"Focus mismatch" (409) when typing

The foreground window changed between the focus call and the type call. Solution: always pass expect_hwnd and verify focus before typing.

Window not found

The window may have been closed or may be a system window that EnumWindows doesn't expose. Try:

curl http://127.0.0.1:8745/api/windows -H "X-Auth-Token: $TOKEN"

Game mode cursor stuck

Use POST /api/release_all or press Esc / Alt+Tab physically.

High OCR latency

OCR on a full 1920×1080 screen can take from a few seconds up to ~30 s depending on your CPU and on-screen complexity. Use a region — small crops are typically 10× faster:

{"region": [0, 0, 800, 100]}

Turkish characters not appearing

The system uses SendInput + KEYEVENTF_UNICODE which is layout-independent. If characters still don't appear, the target app may not support Unicode input — try POST /api/window/post with action: "type" instead.

Project Structure

screen-control/
├── server.py            # Flask HTTP server + all API endpoints
├── core/                # PlatformBackend interface + standardized errors
│   ├── backends.py      # Abstract backend + lazy discovery
│   └── errors.py        # ApiError envelope + error codes
├── backends/            # OS implementations behind PlatformBackend
│   ├── windows.py       # Reference backend (moved from control.py)
│   ├── linux.py         # Fail-closed stub (ROADMAP Phase 5)
│   ├── macos.py         # Fail-closed stub (ROADMAP Phase 7)
│   ├── fake.py          # In-memory backend for tests
│   └── forbidden.py     # Shared blocked-key policy
├── control.py           # Compatibility shim re-exporting backends.windows
├── tests/unit/          # Offline unit + integration tests (no real input)
├── mcp_server.py        # MCP server (stdio + streamable-HTTP) — thin wrapper over the API
├── sdk/
│   └── screen_control.py  # Python SDK client (pip-installable style)
├── index.html           # Bundled web UI (live view + control panels)
├── docs/images/         # README assets (demo GIF captured by the API itself)
├── requirements.txt     # Python dependencies
├── start-server.bat     # One command: REST + MCP + cloud tunnel (Windows)
├── stop-server.bat      # Stop all three processes
├── test-security.py     # Security + game-mode test suite (34 checks)
├── test-game.py         # Live game-mechanics test (app launch → draw → safe close)
├── test-endtoend.py     # End-to-end test: open Notepad → type → save → verify
├── .github/workflows/   # CI: runs the security suite on every push
├── AGENT_GUIDE.md       # AI agent integration guide (separate from this file)
├── README.md            # This file
└── .token               # Auto-generated auth token (gitignored)

Testing

Prerequisites

The server must be running:

cd screen-control
python server.py

Security test suite

Tests authentication, blocked key combos, window management, safe close, critical process protection, and game mode — all non-destructive.

cd screen-control
python test-security.py

Expected output:

== Token Authentication ==
✓ Missing token -> 401
✓ Wrong token -> 401
✓ Correct token -> 200
✓ Non-JSON POST -> 415

== Blocked Key Combos ==
✓ Alt+F4 blocked (403)
✓ Win key blocked (403)
✓ Win+D blocked (403)
✓ Delete blocked (403)

== Window List ==
✓ Windows list requires GET
✓ Window list is non-empty  — 8 windows
✓ Exactly one focused window

== Safe Close Verification ==
✓ Wrong title aborts close

== Critical Process Protection ==
✓ System process (pid 4) rejected (403)
✓ pid 0 rejected (403)

== Game Mode ==
✓ Game mode started
✓ Relative camera look
✓ Game mode stopped

== Watchdog (dry run) ==
✓ Held state returns ok
✓ Watchdog count reported

========================================
RESULT: 19 passed, 0 failed

Live game-mechanics test

Launches a real application (mspaint or notepad), performs hold-to-draw game mechanics, verifies via pixel analysis, then safely closes with "Don't Save" dialog handling.

cd screen-control
python test-game.py

Note: This test launches a real application. It handles cleanup automatically (sends WM_CLOSE and clicks "Don't Save" if a dialog appears).

Contributing

Fork the repository
Create a feature branch
Make your changes
Test on a Windows machine
Submit a pull request

Code Style

Python: PEP 8, type hints, docstrings on all public functions
Docstrings: English, Google style
Error messages: English, descriptive
Comments: English, explain why not what

License

MIT License. See LICENSE for details.

Built with ❤️ for local automation and AI agent research.

Screen Control

Tài liệu

🖥️ Screen Control

Why Screen Control?

Performance Is Agent-Bound

Table of Contents

Features

Architecture

Coordinates & Concurrency

Live-Loop Working Principle

Installation

Requirements

System Requirements

Quick Start

API Reference

Capabilities

Platform Support Matrix

Authentication

Screen Capture

GET /api/screenshot

GET /api/info

Mouse Control

POST /api/mouse

Keyboard Control

POST /api/key

OCR (Screen Reading)

POST /api/ocr

Vision Access (Image Models)

GET /api/vision/frame

GET /api/stream

POST /api/vision/diff

Window Management

GET /api/windows

POST /api/window

Focus-Free (Background) Control

GET /api/window/capture

POST /api/window/post

GET /api/window/input-mode

GET /api/window/children

Virtual Desktops

GET /api/desktops

POST /api/desktop

Game Mode

Safety Endpoints

GET /api/held

POST /api/release_all

🤖 For AI Agents

MCP Support (One-Click Cloud Agents)

How it works

Setup

Desktop agents (stdio transport)

Remote / cloud agents (streamable-HTTP transport)

Headerless connectors (scoped keys)

One-Command Startup (launcher + auto-tunnel)

Available tools (16)

Which transport for whom

Security Model

Threat: Malicious Web Pages (CSRF)

Threat: Stuck Keys / Game Mode Lock

Threat: Wrong Window Typing

Threat: Dangerous Key Combos

Threat: Killing System Processes

Threat: Resource Exhaustion (rogue agent / DoS)

Network Access

Game Mode Guide

Setup

Camera Look

Movement

Minecraft-Specific

Suitability

Vision Access Guide

For Image-Capable Models

For Text-Only Models

Bandwidth Optimization

Troubleshooting

"OCR engine not installed"

Server won't start (port in use)

"Focus mismatch" (409) when typing

Window not found

Game mode cursor stuck

`GET /api/screenshot`

`GET /api/info`

`POST /api/mouse`

`POST /api/key`

`POST /api/ocr`

`GET /api/vision/frame`

`GET /api/stream`

`POST /api/vision/diff`

`GET /api/windows`

`POST /api/window`

`GET /api/window/capture`

`POST /api/window/post`

`GET /api/window/input-mode`

`GET /api/window/children`

`GET /api/desktops`

`POST /api/desktop`

`GET /api/held`

`POST /api/release_all`