faze79/WPFVisualTreeMcp

Overview

Lets AI agents inspect, debug, and control a running WPF app over the Model Context Protocol. Exposes the visual tree, bindings, screenshots, and supports clicks and text input, like Snoop plus Playwright for AI.

Best for

Best for
Developers building AI tools to debug or drive desktop WPF apps

Use cases

• Inspect the live visual tree of a WPF app
• Debug binding values at runtime
• Automate clicks and text input via MCP

How to use

Tools exposed

• wpf_list_processes
• wpf_attach
• wpf_get_visual_tree
• wpf_get_element_properties
• wpf_find_elements
• wpf_find_elements_deep
• wpf_capture_screenshot
• wpf_get_bindings
• wpf_get_binding_errors
• wpf_clear_binding_errors
• wpf_get_data_context
• wpf_get_resources
• wpf_get_styles
• wpf_watch_property
• wpf_evaluate_binding
• wpf_explain_triggers
• wpf_snapshot
• wpf_diff
• wpf_set_property
• wpf_revert_property

Tested with

Claude Code, Cursor, VS Code

Example client config

{\n     "mcpServers": {\n       "wpf-visual-tree": {\n         "command": "C:/path/to/WpfVisualTreeMcp.Server.exe",\n         "args": []\n       }\n     }\n   }

Notes

Lets AI agents inspect, debug, and control a running WPF app over the Model Context Protocol. Exposes the visual tree, bindings, screenshots, and supports clicks and text input, like Snoop plus Playwright for AI.

13 stars on GitHub. Last updated 2026-07-23. Licensed MIT.

Use cases

• Inspect the live visual tree of a WPF app
• Debug binding values at runtime
• Automate clicks and text input via MCP

Pros

• Combines introspection with UI automation over MCP
• Works with standard MCP clients
• No app code changes required

Cons

• Limited to WPF apps (Windows only)
• Low star count, early-stage project
• Requires an MCP client to operate

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.