windows-mcp 0.5.2__py3-none-any.whl → 0.5.3__py3-none-any.whl

main.py → windows_mcp/__main__.py

@@ -1,7 +1,7 @@
 from live_inspect.watch_cursor import WatchCursor
 from contextlib import asynccontextmanager
 from fastmcp.utilities.types import Image
-from src.desktop.service import Desktop
+from windows_mcp.desktop.service import Desktop
 from mcp.types import ToolAnnotations
 from humancursor import SystemCursor
 from textwrap import dedent
@@ -19,6 +19,7 @@ cursor=SystemCursor()
 watch_cursor=WatchCursor()
 windows_version=desktop.get_windows_version()
 default_language=desktop.get_default_language()
+screen_width,screen_height=desktop.get_resolution()
 
 instructions=dedent(f'''
 Windows MCP server provides tools to interact directly with the {windows_version} desktop, 
@@ -39,11 +40,11 @@ mcp=FastMCP(name='windows-mcp',instructions=instructions,lifespan=lifespan)
 
 @mcp.tool(
     name="App-Tool",
-    description="Manages Windows applications through launch, resize, and window switching operations.",
+    description="Manages Windows applications with three modes: 'launch' (start app by name), 'resize' (set window position/size using window_loc=[x,y] and window_size=[width,height]), 'switch' (activate app by name). Essential for application lifecycle management.",
     annotations=ToolAnnotations(
         title="App Tool",
         readOnlyHint=False,
-        destructiveHint=False,
+        destructiveHint=True,
         idempotentHint=False,
         openWorldHint=False
     )
@@ -53,7 +54,7 @@ def app_tool(mode:Literal['launch','resize','switch'],name:str|None=None,window_
     
 @mcp.tool(
     name='Powershell-Tool',
-    description='Execute PowerShell commands and return the output with status code',
+    description='Execute PowerShell commands directly on the Windows system and return output with status code. Supports all PowerShell cmdlets, scripts, and system commands. Use for file operations, system queries, and administrative tasks.',
     annotations=ToolAnnotations(
         title="Powershell Tool",
         readOnlyHint=False,
@@ -68,15 +69,23 @@ def powershell_tool(command: str) -> str:
 
 @mcp.tool(
     name='State-Tool',
-    description='Capture comprehensive desktop state including default language used by user interface, focused/opened applications, interactive UI elements (buttons, text fields, menus), informative content (text, labels, status), and scrollable areas. Optionally includes visual screenshot when use_vision=True. Essential for understanding current desktop context and available UI interactions.',
+    description='Captures complete desktop state including: system language, focused/opened apps, interactive elements (buttons, text fields, links, menus with coordinates), and scrollable areas. Set use_vision=True to include screenshot. Set use_dom=True for browser content to get web page elements instead of browser UI. Always call this first to understand the current desktop state before taking actions.',
     annotations=ToolAnnotations(
         title="State Tool",
         readOnlyHint=True,
+        destructiveHint=False,
+        idempotentHint=True,
         openWorldHint=False
     )
     )
-def state_tool(use_vision:bool=False):
-    desktop_state=desktop.get_state(use_vision=use_vision,as_bytes=True)
+def state_tool(use_vision:bool=False,use_dom:bool=False):
+    # Calculate scale factor to cap resolution at 1080p (1920x1080)
+    max_width, max_height = 1920, 1080
+    scale_width = max_width / screen_width if screen_width > max_width else 1.0
+    scale_height = max_height / screen_height if screen_height > max_height else 1.0
+    scale = min(scale_width, scale_height)  # Use the smaller scale to ensure both dimensions fit
+    
+    desktop_state=desktop.get_state(use_vision=use_vision,use_dom=use_dom,as_bytes=True,scale=scale)
     interactive_elements=desktop_state.tree_state.interactive_elements_to_string()
     scrollable_elements=desktop_state.tree_state.scrollable_elements_to_string()
     apps=desktop_state.apps_to_string()
@@ -100,7 +109,7 @@ def state_tool(use_vision:bool=False):
 
 @mcp.tool(
     name='Click-Tool',
-    description='Click on UI elements at specific coordinates. Supports left/right/middle mouse buttons and single/double/triple clicks. Use coordinates from State-Tool output.',
+    description='Performs mouse clicks at specified coordinates [x, y]. Supports button types: left (default), right (context menu), middle. Supports clicks: 1 (single), 2 (double), 3 (triple). Always use coordinates from State-Tool output to ensure accuracy.',
     annotations=ToolAnnotations(
         title="Click Tool",
         readOnlyHint=False,
@@ -119,7 +128,7 @@ def click_tool(loc:list[int],button:Literal['left','right','middle']='left',clic
 
 @mcp.tool(
     name='Type-Tool',
-    description='Type text into input fields, text areas, or focused elements. Set clear=True to replace existing text, False to append. Click on target element coordinates first.',
+    description='Types text at specified coordinates [x, y]. Set clear=True to clear existing text first (Ctrl+A then type), clear=False to append. Set press_enter=True to submit after typing. Always click on the target input field first to ensure focus.',
     annotations=ToolAnnotations(
         title="Type Tool",
         readOnlyHint=False,
@@ -137,7 +146,7 @@ def type_tool(loc:list[int],text:str,clear:bool=False,press_enter:bool=False)->s
 
 @mcp.tool(
     name='Scroll-Tool',
-    description='Scroll at specific coordinates or current mouse position. Use wheel_times to control scroll amount (1 wheel = ~3-5 lines). Essential for navigating lists, web pages, and long content.',
+    description='Scrolls at coordinates [x, y] or current mouse position if loc=None. Type: vertical (default) or horizontal. Direction: up/down for vertical, left/right for horizontal. wheel_times controls amount (1 wheel ≈ 3-5 lines). Use for navigating long content, lists, and web pages.',
     annotations=ToolAnnotations(
         title="Scroll Tool",
         readOnlyHint=False,
@@ -156,7 +165,7 @@ def scroll_tool(loc:list[int]=None,type:Literal['horizontal','vertical']='vertic
 
 @mcp.tool(
     name='Drag-Tool',
-    description='Drag and drop operation from current coordinates to destination coordinates. Useful for moving files, resizing windows, or drag-and-drop interactions.',
+    description='Performs drag-and-drop from current mouse position to destination coordinates [x, y]. Click or move to source position first, then call this tool with target coordinates. Use for moving files, reordering items, resizing windows, or any drag-drop UI interactions.',
     annotations=ToolAnnotations(
         title="Drag Tool",
         readOnlyHint=False,
@@ -174,7 +183,7 @@ def drag_tool(to_loc:list[int])->str:
 
 @mcp.tool(
     name='Move-Tool',
-    description='Move mouse cursor to specific coordinates without clicking. Useful for hovering over elements or positioning cursor before other actions.',
+    description='Moves mouse cursor to coordinates [x, y] without clicking. Use for hovering to reveal tooltips/menus, positioning cursor before drag operations, or triggering hover-based UI changes. Does not interact with elements.',
     annotations=ToolAnnotations(
         title="Move Tool",
         readOnlyHint=False,
@@ -192,7 +201,7 @@ def move_tool(to_loc:list[int])->str:
 
 @mcp.tool(
     name='Shortcut-Tool',
-    description='Execute keyboard shortcuts using key combinations. Pass keys as list (e.g., ctrl+c for copy, alt+tab for app switching, win+r for Run dialog, win is for opening the start menu).',
+    description='Executes keyboard shortcuts using key combinations separated by +. Examples: "ctrl+c" (copy), "ctrl+v" (paste), "alt+tab" (switch apps), "win+r" (Run dialog), "win" (Start menu), "ctrl+shift+esc" (Task Manager). Use for quick actions and system commands.',
     annotations=ToolAnnotations(
         title="Shortcut Tool",
         readOnlyHint=False,
@@ -207,10 +216,12 @@ def shortcut_tool(shortcut:str):
 
 @mcp.tool(
     name='Wait-Tool',
-    description='Pause execution for specified duration in seconds. Useful for waiting for applications to load, animations to complete, or adding delays between actions.',
+    description='Pauses execution for specified duration in seconds. Use when waiting for: applications to launch/load, UI animations to complete, page content to render, dialogs to appear, or between rapid actions. Helps ensure UI is ready before next interaction.',
     annotations=ToolAnnotations(
         title="Wait Tool",
         readOnlyHint=True,
+        destructiveHint=False,
+        idempotentHint=True,
         openWorldHint=False
     )
     )
@@ -220,16 +231,26 @@ def wait_tool(duration:int)->str:
 
 @mcp.tool(
     name='Scrape-Tool',
-    description='Fetch and convert webpage content to markdown format. Provide full URL including protocol (http/https). Returns structured text content suitable for analysis.',
+    description='Extracts visible text content from the currently focused browser tab. Returns content in plain text format with scroll status indicators (top/bottom reached or more content available). Only works when a browser with DOM is active. Use State-Tool with use_dom=True first to ensure browser is ready.',
     annotations=ToolAnnotations(
         title="Scrape Tool",
         readOnlyHint=True,
+        destructiveHint=False,
+        idempotentHint=True,
         openWorldHint=True
     )
     )
 def scrape_tool(url:str)->str:
-    content=desktop.scrape(url)
-    return f'Scraped the contents of the entire webpage:\n{content}'
+    desktop_state=desktop.desktop_state
+    tree_state=desktop_state.tree_state
+    if not tree_state.dom_node:
+        return f'Unable to scrape URL: {url}. No DOM node found.'
+    dom_node=tree_state.dom
+    vertical_scroll_percent=dom_node.vertical_scroll_percent
+    content='\n'.join([node.text for node in tree_state.dom_informative_nodes])
+    header_status = "Reached top" if vertical_scroll_percent <= 0 else "Scroll up to see more"
+    footer_status = "Reached bottom" if vertical_scroll_percent >= 100 else "Scroll down to see more"
+    return f'URL:{url}\nContent:\n{header_status}\n{content}\n{footer_status}'
 
 
 @click.command()

{src → windows_mcp}/desktop/service.py

@@ -1,10 +1,10 @@
-from src.desktop.config import BROWSER_NAMES, PROCESS_PER_MONITOR_DPI_AWARE
-from src.desktop.views import DesktopState, App, Size, Status
+from windows_mcp.desktop.config import BROWSER_NAMES, PROCESS_PER_MONITOR_DPI_AWARE
+from windows_mcp.desktop.views import DesktopState, App, Size, Status
 from locale import getpreferredencoding
 from contextlib import contextmanager
 from typing import Optional,Literal
 from markdownify import markdownify
-from src.tree.service import Tree
+from windows_mcp.tree.service import Tree
 from fuzzywuzzy import process
 from psutil import Process
 from time import sleep
@@ -46,7 +46,10 @@ class Desktop:
         self.tree=Tree(self)
         self.desktop_state=None
         
-    def get_state(self,use_vision:bool=False,as_bytes:bool=False)->DesktopState:
+    def get_resolution(self)->tuple[int,int]:
+        return pg.size()
+        
+    def get_state(self,use_vision:bool=False,use_dom:bool=False,as_bytes:bool=False,scale:float=1.0)->DesktopState:
         sleep(0.1)
         apps=self.get_apps()
         active_app=self.get_active_app()
@@ -54,9 +57,9 @@ class Desktop:
             apps.remove(active_app)
         logger.debug(f"Active app: {active_app}")
         logger.debug(f"Apps: {apps}")
-        tree_state=self.tree.get_state(active_app,apps)
+        tree_state=self.tree.get_state(active_app,apps,use_dom=use_dom)
         if use_vision:
-            screenshot=self.tree.annotated_screenshot(tree_state.interactive_nodes)
+            screenshot=self.tree.get_annotated_screenshot(tree_state.interactive_nodes,scale=scale)
             if as_bytes:
                 bytes_io=io.BytesIO()
                 screenshot.save(bytes_io,format='PNG')

{src → windows_mcp}/desktop/views.py

@@ -1,4 +1,4 @@
-from src.tree.views import TreeState
+from windows_mcp.tree.views import TreeState
 from dataclasses import dataclass
 from tabulate import tabulate
 from typing import Optional

{src → windows_mcp}/tree/service.py

@@ -1,11 +1,11 @@
-from src.tree.config import INTERACTIVE_CONTROL_TYPE_NAMES,DOCUMENT_CONTROL_TYPE_NAMES,INFORMATIVE_CONTROL_TYPE_NAMES, DEFAULT_ACTIONS, THREAD_MAX_RETRIES
+from windows_mcp.tree.config import INTERACTIVE_CONTROL_TYPE_NAMES,DOCUMENT_CONTROL_TYPE_NAMES,INFORMATIVE_CONTROL_TYPE_NAMES, DEFAULT_ACTIONS, THREAD_MAX_RETRIES
 from uiautomation import Control,ImageControl,ScrollPattern,WindowControl,Rect,GetRootControl,PatternId
-from src.tree.views import TreeElementNode, ScrollElementNode, Center, BoundingBox, TreeState
+from windows_mcp.tree.views import TreeElementNode, ScrollElementNode, TextElementNode, Center, BoundingBox, TreeState
 from concurrent.futures import ThreadPoolExecutor, as_completed
-from src.tree.utils import random_point_within_bounding_box
+from windows_mcp.tree.utils import random_point_within_bounding_box
 from PIL import Image, ImageFont, ImageDraw
-from src.desktop.views import App
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING,Optional
+from windows_mcp.desktop.views import App
 from time import sleep
 import logging
 import random
@@ -18,36 +18,74 @@ handler.setFormatter(formatter)
 logger.addHandler(handler)
 
 if TYPE_CHECKING:
-    from src.desktop.service import Desktop
+    from windows_mcp.desktop.service import Desktop
     
 class Tree:
     def __init__(self,desktop:'Desktop'):
         self.desktop=desktop
-        screen_size=self.desktop.get_screen_size()
+        self.screen_size=self.desktop.get_screen_size()
+        self.dom:Optional[Control]=None
         self.dom_bounding_box:BoundingBox=None
         self.screen_box=BoundingBox(
-            top=0, left=0, bottom=screen_size.height, right=screen_size.width,
-            width=screen_size.width, height=screen_size.height 
+            top=0, left=0, bottom=self.screen_size.height, right=self.screen_size.width,
+            width=self.screen_size.width, height=self.screen_size.height 
         )
+        self.root:Optional[TreeElementNode]=None
 
-    def get_state(self,active_app:App,other_apps:list[App])->TreeState:
-        root=GetRootControl()
+    def get_state(self,active_app:App,other_apps:list[App],use_dom:bool=False)->TreeState:
+        self.root=GetRootControl()
         other_apps_handle=set(map(lambda other_app: other_app.handle,other_apps))
-        apps=list(filter(lambda app:app.NativeWindowHandle not in other_apps_handle,root.GetChildren()))
+        apps=list(filter(lambda app:app.NativeWindowHandle not in other_apps_handle,self.root.GetChildren()))
         del other_apps_handle
         if active_app:
             apps=list(filter(lambda app:app.ClassName!='Progman',apps))
-        interactive_nodes,scrollable_nodes=self.get_appwise_nodes(apps=apps)
-        return TreeState(interactive_nodes=interactive_nodes,scrollable_nodes=scrollable_nodes)
+        interactive_nodes,scrollable_nodes,dom_informative_nodes=self.get_appwise_nodes(apps=apps,use_dom=use_dom)
+        root=TreeElementNode(**{
+            'name':'Desktop',
+            'control_type':'PaneControl',
+            'app_name':'Desktop',
+            'value':'',
+            'shortcut':'',
+            'bounding_box':self.screen_box,
+            'center':Center(x=self.screen_box.left+self.screen_box.width//2,y=self.screen_box.top+self.screen_box.height//2),
+            'xpath':'',
+            'is_focused':False
+        })
+        dom=None
+        if self.dom:
+            scroll_pattern=self.dom.GetPattern(PatternId.ScrollPattern)
+            bounding_box=self.dom.BoundingRectangle
+            dom=ScrollElementNode(**{
+                'name':"DOM",
+                'control_type':'DocumentControl',
+                'app_name':"DOM",
+                'bounding_box':BoundingBox(
+                    left=bounding_box.left,
+                    top=bounding_box.top,
+                    right=bounding_box.right,
+                    bottom=bounding_box.bottom,
+                    width=bounding_box.width(),
+                    height=bounding_box.height()
+                ),
+                'center':Center(x=bounding_box.left+bounding_box.width()//2,y=bounding_box.top+bounding_box.height()//2),
+                'horizontal_scrollable':scroll_pattern.HorizontallyScrollable,
+                'horizontal_scroll_percent':scroll_pattern.HorizontalScrollPercent if scroll_pattern.HorizontallyScrollable else 0,
+                'vertical_scrollable':scroll_pattern.VerticallyScrollable,
+                'vertical_scroll_percent':scroll_pattern.VerticalScrollPercent if scroll_pattern.VerticallyScrollable else 0,
+                'xpath':'',
+                'is_focused':False
+            })
+        return TreeState(root=root,dom=dom,interactive_nodes=interactive_nodes,scrollable_nodes=scrollable_nodes,dom_informative_nodes=dom_informative_nodes)
 
-    def get_appwise_nodes(self,apps:list[Control]) -> tuple[list[TreeElementNode],list[ScrollElementNode]]:
-        interactive_nodes, scrollable_nodes = [], []
+    def get_appwise_nodes(self,apps:list[Control],use_dom:bool=False)-> tuple[list[TreeElementNode],list[ScrollElementNode],list[TextElementNode]]:
+        interactive_nodes, scrollable_nodes,dom_informative_nodes = [], [], []
         with ThreadPoolExecutor() as executor:
             retry_counts = {app: 0 for app in apps}
             future_to_app = {
                 executor.submit(
                     self.get_nodes, app, 
-                    self.desktop.is_app_browser(app)
+                    self.desktop.is_app_browser(app),
+                    use_dom
                 ): app 
                 for app in apps
             }
@@ -57,18 +95,20 @@ class Tree:
                     try:
                         result = future.result()
                         if result:
-                            element_nodes, scroll_nodes = result
+                            element_nodes, scroll_nodes,informative_nodes = result
                             interactive_nodes.extend(element_nodes)
                             scrollable_nodes.extend(scroll_nodes)
+                            dom_informative_nodes.extend(informative_nodes)
                     except Exception as e:
                         retry_counts[app] += 1
                         logger.debug(f"Error in processing node {app.Name}, retry attempt {retry_counts[app]}\nError: {e}")
                         if retry_counts[app] < THREAD_MAX_RETRIES:
-                            new_future = executor.submit(self.get_nodes, app, self.desktop.is_app_browser(app))
+                            logger.debug(f"Retrying {app.Name} for the {retry_counts[app]}th time")
+                            new_future = executor.submit(self.get_nodes, app, self.desktop.is_app_browser(app),use_dom)
                             future_to_app[new_future] = app
                         else:
                             logger.error(f"Task failed completely for {app.Name} after {THREAD_MAX_RETRIES} retries")
-        return interactive_nodes,scrollable_nodes
+        return interactive_nodes,scrollable_nodes,dom_informative_nodes
     
     def iou_bounding_box(self,window_box: Rect,element_box: Rect,) -> BoundingBox:
         # Step 1: Intersection of element and window (existing logic)
@@ -105,7 +145,7 @@ class Tree:
             )
         return bounding_box
 
-    def get_nodes(self, node: Control, is_browser:bool=False) -> tuple[list[TreeElementNode],list[ScrollElementNode]]:
+    def get_nodes(self, node: Control, is_browser:bool=False,use_dom:bool=False) -> tuple[list[TreeElementNode],list[ScrollElementNode]]:
         window_bounding_box=node.BoundingRectangle
 
         def is_element_visible(node:Control,threshold:int=0):
@@ -331,11 +371,10 @@ class Tree:
                         'is_focused':is_focused
                     })
                     interactive_nodes.append(tree_node)
-            # elif is_element_text(node):
-            #     informative_nodes.append(TextElementNode(
-            #         name=node.Name.strip() or "''",
-            #         app_name=app_name
-            #     ))
+            elif is_element_text(node):
+                dom_informative_nodes.append(TextElementNode(
+                    text=node.Name.strip(),
+                ))
             
             children=node.GetChildren()
 
@@ -344,11 +383,12 @@ class Tree:
                 # Incrementally building the xpath
                 
                 # Check if the child is a DOM element
-                if is_browser and child.ClassName == "Chrome_RenderWidgetHostHWND":
+                if is_browser and child.AutomationId == "RootWebArea":
                     bounding_box=child.BoundingRectangle
                     self.dom_bounding_box=BoundingBox(left=bounding_box.left,top=bounding_box.top,
                     right=bounding_box.right,bottom=bounding_box.bottom,width=bounding_box.width(),
                     height=bounding_box.height())
+                    self.dom=child
                     # enter DOM subtree
                     tree_traversal(child, is_dom=True, is_dialog=is_dialog)
                 # Check if the child is a dialog
@@ -369,7 +409,7 @@ class Tree:
                     # normal non-dialog children
                     tree_traversal(child, is_dom=is_dom, is_dialog=is_dialog)
 
-        interactive_nodes, dom_interactive_nodes, scrollable_nodes = [], [], []
+        interactive_nodes, dom_interactive_nodes, scrollable_nodes, dom_informative_nodes = [], [], [], []
         app_name=node.Name.strip()
         match node.ClassName:
             case "Progman":
@@ -386,12 +426,25 @@ class Tree:
         logger.debug(f'DOM interactive nodes:{len(dom_interactive_nodes)}')
         logger.debug(f'Scrollable nodes:{len(scrollable_nodes)}')
 
-        interactive_nodes.extend(dom_interactive_nodes)
-        return (interactive_nodes,scrollable_nodes)
+        if use_dom:
+            if is_browser:
+                return (dom_interactive_nodes,scrollable_nodes,dom_informative_nodes)
+            else:
+                return ([],[],[])
+        else:
+            return (interactive_nodes+dom_interactive_nodes,scrollable_nodes,dom_informative_nodes)
 
-    def annotated_screenshot(self, nodes: list[TreeElementNode]) -> Image.Image:
+    def get_annotated_screenshot(self, nodes: list[TreeElementNode],scale:float=1.0) -> Image.Image:
         screenshot = self.desktop.get_screenshot()
         sleep(0.10)
+        
+        original_width = screenshot.width
+        original_height = screenshot.height
+
+        scaled_width = int(original_width * scale)
+        scaled_height = int(original_height * scale)
+        screenshot = screenshot.resize((scaled_width, scaled_height), Image.Resampling.LANCZOS)
+        
         # Add padding
         padding = 5
         width = int(screenshot.width + (1.5 * padding))
@@ -413,12 +466,12 @@ class Tree:
             box = node.bounding_box
             color = get_random_color()
 
-            # Scale and pad the bounding box also clip the bounding box
+            # Scale and pad the bounding box coordinates
             adjusted_box = (
-                int(box.left) + padding,
-                int(box.top) + padding,
-                int(box.right) + padding,
-                int(box.bottom) + padding
+                int(box.left * scale) + padding,
+                int(box.top * scale) + padding,
+                int(box.right * scale) + padding,
+                int(box.bottom * scale) + padding
             )
             # Draw bounding box
             draw.rectangle(adjusted_box, outline=color, width=2)

{src → windows_mcp}/tree/views.py

@@ -1,10 +1,14 @@
 from dataclasses import dataclass,field
 from tabulate import tabulate
+from typing import Optional
 
 @dataclass
 class TreeState:
+    root:Optional['TreeElementNode']=None
+    dom:Optional['ScrollElementNode']=None
     interactive_nodes:list['TreeElementNode']=field(default_factory=list)
     scrollable_nodes:list['ScrollElementNode']=field(default_factory=list)
+    dom_informative_nodes:list['TextElementNode']=field(default_factory=list)
 
     def interactive_elements_to_string(self) -> str:
         if not self.interactive_nodes:
@@ -99,4 +103,8 @@ class ScrollElementNode:
             self.is_focused
         ]
 
+@dataclass
+class TextElementNode:
+    text:str
+
 ElementNode=TreeElementNode|ScrollElementNode

{windows_mcp-0.5.2.dist-info → windows_mcp-0.5.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: windows-mcp
-Version: 0.5.2
+Version: 0.5.3
 Summary: Lightweight MCP Server for interacting with Windows Operating System.
 Project-URL: homepage, https://github.com/CursorTouch
 Author-email: Jeomon George <jeogeoalukka@gmail.com>
@@ -44,7 +44,6 @@ Requires-Dist: python-levenshtein>=0.27.1
 Requires-Dist: pywinauto>=0.6.9
 Requires-Dist: requests>=2.32.3
 Requires-Dist: tabulate>=0.9.0
-Requires-Dist: twine>=6.2.0
 Requires-Dist: uiautomation>=2.0.24
 Description-Content-Type: text/markdown
 
@@ -114,6 +113,58 @@ mcp-name: io.github.CursorTouch/Windows-MCP
 - **Real-Time Interaction**  
   Typical latency between actions (e.g., from one mouse click to the next) ranges from **0.7 to 2.5 secs**, and may slightly vary based on the number of active applications and system load, also the inferencing speed of the llm.
 
+- **DOM Mode for Browser Automation**  
+  Special `use_dom=True` mode for State-Tool that focuses exclusively on web page content, filtering out browser UI elements for cleaner, more efficient web automation.
+
+## 🌐 DOM Mode for Browser Automation
+
+Windows-MCP includes a powerful **DOM Mode** feature that enhances browser automation by focusing on web page content rather than browser UI elements.
+
+### What is DOM Mode?
+
+When `use_dom=True` is set in the State-Tool, the MCP server:
+- **Filters out browser UI**: Removes address bars, tabs, toolbars, and other browser chrome elements
+- **Returns only web content**: Provides interactive elements (links, buttons, forms) from the actual web page
+- **Reduces token usage**: Cleaner output means fewer tokens sent to the LLM
+- **Improves accuracy**: LLM focuses only on relevant web page elements
+
+### When to Use DOM Mode
+
+✅ **Use `use_dom=True` when:**
+- Automating web applications or websites
+- Scraping web content
+- Filling out web forms
+- Clicking links or buttons on web pages
+- Testing web interfaces
+- You want to ignore browser UI and focus on page content
+
+❌ **Use `use_dom=False` (default) when:**
+- Interacting with browser controls (address bar, tabs, bookmarks)
+- Working with desktop applications
+- Need to see all UI elements including browser chrome
+- Managing browser settings or extensions
+
+### Example Usage
+
+```python
+# Get web page content only (no browser UI)
+state_tool(use_vision=False, use_dom=True)
+
+# Get full desktop state including browser UI
+state_tool(use_vision=False, use_dom=False)
+
+# Get web page content with screenshot
+state_tool(use_vision=True, use_dom=True)
+```
+
+### Benefits
+
+1. **Token Efficiency**: Reduces the amount of data sent to LLM by filtering irrelevant browser UI
+2. **Better Focus**: LLM concentrates on actionable web page elements
+3. **Cleaner Output**: Only relevant interactive elements from the DOM are returned
+4. **Faster Processing**: Less data means faster LLM inference
+5. **Cost Savings**: Fewer tokens = lower API costs for cloud LLMs
+
 ## 🛠️Installation
 
 ### Prerequisites
@@ -317,7 +368,7 @@ MCP Client can access the following tools to interact with Windows:
 - `Move-Tool`: Move mouse pointer.
 - `Shortcut-Tool`: Press keyboard shortcuts (`Ctrl+c`, `Alt+Tab`, etc).
 - `Wait-Tool`: Pause for a defined duration.
-- `State-Tool`: Combined snapshot of default language, browser, active apps and interactive, textual and scrollable elements along with screenshot of the desktop..
+- `State-Tool`: Combined snapshot of default language, browser, active apps and interactive, textual and scrollable elements along with screenshot of the desktop. Supports `use_dom=True` for browser content extraction (web page elements only) and `use_vision=True` for including screenshots.
 - `App-Tool`: To launch an application from the start menu, resize or move the window and switch between apps.
 - `Shell-Tool`: To execute PowerShell commands.
 - `Scrape-Tool`: To scrape the entire webpage for information.

windows_mcp-0.5.3.dist-info/RECORD

@@ -0,0 +1,16 @@
+windows_mcp/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+windows_mcp/__main__.py,sha256=kCgkB5ckRlb7hgjg_Gpj_OQWiWJdgEOEMcBFJ7Kqmy8,11920
+windows_mcp/desktop/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+windows_mcp/desktop/config.py,sha256=7rAb64pmC275PpNRXVOyOf0Psu089AOosRC8T5kVGWA,384
+windows_mcp/desktop/service.py,sha256=97e2E4TdMs3TwW6CtupVxnwhWqdBKU5eH4MDz6_5Hmk,18469
+windows_mcp/desktop/views.py,sha256=_hZ5sfY1uWVi5mpaysVd-plwP_DT6SXpKa33Z8WT6gI,1523
+windows_mcp/tree/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+windows_mcp/tree/config.py,sha256=k-Mjo_yIn0d1AzcEW_bxiaXyBFxBZZSyy7hCNQ3XVp0,1010
+windows_mcp/tree/service.py,sha256=KWdOHY5Q1HU-PJV6vMv_h9KVb5oL_E5vnAW-KICEzfw,24786
+windows_mcp/tree/utils.py,sha256=6hbxdIQPrAY-I3jcHsRqodHlxboTQj2GnLA71bf1lqY,911
+windows_mcp/tree/views.py,sha256=6A1bLGVt_MHPTvQt9kbUFoPpIqMI43JZjOSg-_o3ajk,3479
+windows_mcp-0.5.3.dist-info/METADATA,sha256=aMbmdQu1I-6gK58_Lyz3lA5ZfbzI9XXRUNwNwOaey7o,14541
+windows_mcp-0.5.3.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+windows_mcp-0.5.3.dist-info/entry_points.txt,sha256=wW8NcVQ_OJK5e5GemZSE_nOKyxfUtBPq2acFLszRwaw,58
+windows_mcp-0.5.3.dist-info/licenses/LICENSE.md,sha256=U1UM4Xi_IX-jHnHjGT0rETNia-Ck8gd92iSQMqQ6a8Y,1089
+windows_mcp-0.5.3.dist-info/RECORD,,

{windows_mcp-0.5.2.dist-info → windows_mcp-0.5.3.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: hatchling 1.27.0
+Generator: hatchling 1.28.0
 Root-Is-Purelib: true
 Tag: py3-none-any

windows_mcp-0.5.3.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+windows-mcp = windows_mcp.__main__:main

windows_mcp-0.5.2.dist-info/RECORD

@@ -1,16 +0,0 @@
-main.py,sha256=Bg_iHXmNxIE1uUioBf0OMEolNkYisGCManA9tpLzv5w,9630
-src/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-src/desktop/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-src/desktop/config.py,sha256=7rAb64pmC275PpNRXVOyOf0Psu089AOosRC8T5kVGWA,384
-src/desktop/service.py,sha256=yzB1SFS2h1fSxMHsYOwa0mJLTOSdIyDWAmfex-DX3dM,18295
-src/desktop/views.py,sha256=vDPPUfD8vNkCS_4-vc-bA4tqG-klqDtznypAQJCN4TA,1515
-src/tree/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-src/tree/config.py,sha256=k-Mjo_yIn0d1AzcEW_bxiaXyBFxBZZSyy7hCNQ3XVp0,1010
-src/tree/service.py,sha256=5RIaabVBwmdKSsmaxTV8UW2f6VFwmyeJTvNWhoudTeM,21864
-src/tree/utils.py,sha256=6hbxdIQPrAY-I3jcHsRqodHlxboTQj2GnLA71bf1lqY,911
-src/tree/views.py,sha256=DVgB8x7Mg9NaZL5xZzhOAzgLuwFw6DWFTLK5hIxWsvk,3232
-windows_mcp-0.5.2.dist-info/METADATA,sha256=Vp5YyAirr8qtj7SMekByPhW8Fx9PbpamHVFsD2X1xlY,12380
-windows_mcp-0.5.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-windows_mcp-0.5.2.dist-info/entry_points.txt,sha256=NMSKckn68nbiSSmQ9eFiP8cmPrDSR_vzeYE-Zqmhn_o,42
-windows_mcp-0.5.2.dist-info/licenses/LICENSE.md,sha256=U1UM4Xi_IX-jHnHjGT0rETNia-Ck8gd92iSQMqQ6a8Y,1089
-windows_mcp-0.5.2.dist-info/RECORD,,

windows_mcp-0.5.2.dist-info/entry_points.txt

@@ -1,2 +0,0 @@
-[console_scripts]
-windows-mcp = main:main