mcp -workshop
Mastra MCP Server workshop
claude mcp add --transport stdio mastra-ai-mcp-server-workshop npx -y mastra-ai-mcp-server-workshop \ --env DESCRIPTION="Mastra MCP Server Workshop (local development) - adjust as needed"
How to use
The Mastra MCP Server Workshop provides a local MCP server that serves markdown notes from a local directory, exposes tools for the agent to create and edit those notes, and pairs with a reactive Next.js frontend to present and manipulate the content. The server is designed to illustrate how to build custom agent tools (like writeNote) and prompts, and how to push updates to a client in real time. In practice, you can run the MCP server separately from the frontend and the agent, reflecting a typical production setup where services are decoupled. Once running, you can interact with the agent through slash-command prompts to create notes, edit content, and trigger predefined prompts such as brainstorming ideas or summarizing a note. The frontend offers a list of notes, a note viewer, and an agent panel that uses these server capabilities to manage your local markdown repository.
How to install
Prerequisites:
- Node.js and pnpm installed on your machine
- Access to a terminal/command prompt
Install dependencies (from the repository root):
pnpm install
Run the MCP server (separate from the frontend and agent):
pnpm dev:mcp
Run the Mastra agent (optional, to demonstrate the agent interacting with the MCP server):
pnpm dev:agent
Run the frontend (Next.js app):
pnpm dev:web
Open the application in your browser at http://localhost:3000
Additional notes
Notes:
- The MCP server exposes markdown notes from the local packages/notes directory. Ensure the directory exists and contains your initial notes.
- The agent has a writeNote tool to create or modify notes on the filesystem. Use descriptive titles and content to see updates reflected in real-time.
- The frontend includes a slash-command menu. Type / in the agent chat to bring up prompts like brainstorm_ideas or summarize_note. Prompts operate using the current note context when available.
- In a production workflow, run the MCP server and agent separately; the agent should be configured to reach the MCP server endpoint. This separation mirrors typical service-oriented architectures and improves reliability during development.
- If you modify notes outside of the MCP tools, you may need to refresh the frontend to see changes reflected in the UI.
Related MCP Servers
iterm
A Model Context Protocol server that executes commands in the current iTerm session - useful for REPL and CLI assistance
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
editor
MCP Server for Phaser Editor
DoorDash
MCP server from JordanDalton/DoorDash-MCP-Server
mcp
MCP сервер для автоматического создания и развертывания приложений в Timeweb Cloud