PyWry

Overview

PyWry is a Python-based UI toolkit that enables developers to create native desktop apps, Jupyter widgets, and web applications from a single API. It features comprehensive MCP server functionality with tools for creating interactive widgets, displaying data visualizations, integrating chat interfaces, and building TradingView-style charts. The MCP server exposes tools like create_widget, show_plotly, show_dataframe, and create_chat_widget, delivering HTML artifacts via the Model Context Protocol.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• create_widget

  Create a custom widget with HTML content and toolbar components

  Try:Create a widget with a header and a buttonTry:Build a custom dashboard with toolbar controls
• show_plotly

  Display a Plotly chart as an interactive widget

  Try:Show a scatter plot of the iris datasetTry:Create a bar chart comparing sales data
• show_dataframe

  Render a pandas DataFrame in an interactive AgGrid widget

  Try:Display the DataFrame with filtering capabilitiesTry:Show a table of employee data with selection events
• show_tvchart

  Display an interactive TradingView chart with technical analysis tools

  Try:Show a candlestick chart with volume indicatorsTry:Create a financial chart with drawing tools
• create_chat_widget

  Create a streaming chat widget with threads and artifacts

  Try:Build a chat interface for user supportTry:Create a collaborative discussion space
• get_widget_app

  Get an updated snapshot of an existing widget

  Try:Refresh the widget with new dataTry:Update the chart with the latest values

Comparable tools

Installation

pip install 'pywry[mcp]'

pywry mcp --transport stdio

{
  "mcpServers": {
    "pywry": {
      "command": "pywry",
      "args": ["mcp", "--transport", "stdio"]
    }
  }
}

FAQ

What output formats does PyWry support?

PyWry supports three output targets: native desktop windows via OS webview, Jupyter widgets for notebook environments, and browser tabs as web applications.

How do the MCP server tools work?

Widget-creating tools return an AppArtifact - a self-contained HTML snapshot delivered as an MCP EmbeddedResource with mimeType text/html. Each render bumps a widget revision, keeping a live WebSocket bridge to Python for the latest version.

Compare PyWry with