promptboard
The Shared Whiteboard for Your AI Agents via MCP. Paste screenshots, mark them up, and share with AI.
claude mcp add --transport stdio pjueon-promptboard npx -y promptboard
How to use
PromptBoard is an MCP server that provides a shared visual whiteboard environment for AI agents and users. It allows pasting screenshots directly, marking up visuals, and sharing context across MCP-enabled clients. The server exposes a visual canvas with drawing tools, multi-image pasting via clipboard, and an interface designed to be accessed by MCP-enabled assistants (like Claude Code and Gemini CLI) through the standard MCP workflow. It also supports an analysis command to interpret the whiteboard content via the /check-board command. This makes it easy to collaboratively reason about images, annotations, and visual context within conversations with AI agents.
From common MCP clients, you can configure and invoke PromptBoard using the provided MCP entry points. Typical usage involves adding the promptboard server to your MCP configuration so that clients can launch or attach to the shared whiteboard, then using the built-in commands to open the board or analyze its contents. For example, you can open the whiteboard to visualize context, or issue a check-board query to get a concise interpretation of the current visuals and markings. The server supports keyboard shortcuts and commands that streamline collaboration across teammates and AI agents.
How to install
Prerequisites
- Node.js and npm (or npx) installed on your system
- Internet access to fetch the promptboard package from npm
Installation steps
-
Ensure Node.js and npm are installed. Verify:
node -v npm -v
-
Install and run PromptBoard via npx (no global install required):
npx -y promptboard
-
Integrate with your MCP client by adding the server to your MCP configuration (see mcp_config below).
-
Launch your MCP client and ensure PromptBoard initializes correctly, then open the whiteboard from the client UI or using the provided commands.
Additional notes
Tips and notes:
- The MCP configuration uses the npx approach, so the first run may download the promptboard package. This is expected behavior for on-demand execution.
- You can paste images directly onto the canvas using clipboard paste (Ctrl+V).
- The whiteboard supports basic drawing tools, shapes, text, and an eraser. Use Undo/Redo (Ctrl+Z / Ctrl+Shift+Z) to manage changes.
- The server exposes a /check-board command to analyze the current whiteboard content and provide a contextual interpretation. Integrate this with your MCP client to automate insights from visuals.
- If you modify client configurations for Claude, Gemini, or other MCP tools, maintain consistency by using the same promptboard MCP entry to avoid drift between environments.
Related MCP Servers
Mantic.sh
A structural code search engine for Al agents.
mcp-atlassian
MCP server connecting AI assistants with Jira & Confluence for smart project management.
mcpresso
TypeScript framework to build robust, agent-ready MCP servers around your APIs.
mcp-micromanage-your-agent
A Micromanagement Tool for Development Workflows: Helps coding agent plan, track, and visualize sequential development tasks with detailed commit-level granularity. Features interactive visualization, automated status tracking, and structured workflow management.
mcp-sys-bridge
An implementation of the Model Context Protocol (MCP), acting as a simple bridge to native OS functionalities like clipboard management and URL handling.
dbt-core
dbt Core MCP Server: Interact with dbt projects via Model Context Protocol