Interactive Drawing MCP

What is Interactive Drawing MCP?

Interactive Drawing MCP is an example Model Context Protocol (MCP) server that provides an interactive drawing interface. It integrates a 16×16 grid UI (built with Tkinter) with AI assistants like Goose, allowing users to toggle cells, retrieve grid state, and persist the drawing across sessions.

How to use Interactive Drawing MCP?

Install from source with pip install -e ., then run the server standalone with interactive-draw-mcp or launch it with Goose via goose session --with-extension "python -m interactive-draw-mcp". Tools are available to start a session, toggle a cell, and get the grid state.

Key features of Interactive Drawing MCP

• 16×16 grid with toggleable cells (filled/empty)
• Clear grid button to reset all cells
• Real‑time UI updates via background file watching
• Grid state persisted to board_state.json
• Three server tools: start session, toggle cell, get grid state

Use cases of Interactive Drawing MCP

• Demonstrating an MCP server with a separate graphical UI
• Experimenting with AI‑assisted drawing via Goose
• Learning how to build MCP extensions that include interactive windows

FAQ from Interactive Drawing MCP

What dependencies does Interactive Drawing MCP require?

It requires Python and Tkinter (typically included with Python). The package is installed from source.

How does the grid state persist across sessions?

The grid state is saved to a JSON file (board_state.json) and reloaded on subsequent runs.

Does Interactive Drawing MCP support any transport or authentication?

No. The server uses the standard MCP protocol (stdio transport) with no authentication; it is designed for local development and demonstration.

Can I use Interactive Drawing MCP with AI assistants other than Goose?

The README only mentions integration with Goose. The server exposes standard MCP tools, so it may work with any MCP‑compatible client, but this is not documented.

What is the grid size and how do I toggle cells?

The grid is 16×16. Cells are toggled via the toggle_cell_color(row, col) tool, or by clicking in the UI window.