Fizzy MCP

What is Fizzy MCP?

Fizzy MCP is a Model Context Protocol (MCP) server that connects AI assistants like Claude, Cursor, and OpenCode to Fizzy - the AI-powered task management platform.

With Fizzy MCP, your AI can:

• Create, update, and manage cards and boards
• Triage work into workflow columns
• Assign tasks to team members
• Track progress with tags and golden cards
• Collaborate through comments

40 MCP Tools

Complete API coverage with tools for boards, cards, columns, comments, tags, and users.

Two Deployment Modes

Choose Remote (hosted service) or Local (stdio) based on your needs.

Auto-Configuration

Automatically detect and configure Claude Desktop, Cursor, and OpenCode.

Type-Safe

Built with TypeScript and Zod for reliable, validated API interactions.

Deployment Modes

Fizzy MCP offers two deployment modes to fit your workflow:

Remote Mode (Recommended)

Use our hosted service at fizzy-mcp-hosted.ryanyogan.workers.dev. No local server required - just configure and go. Best for: Quick setup, remote MCP clients, environments where local installation isn’t practical. Rate limits: 100 writes/day (anonymous) or 1,000/day (with API key)

Local Mode

Run the MCP server locally via stdio transport. Full control with no rate limits. Best for: Desktop AI assistants, development, testing, and offline usage.

Quick Install

npx fizzy-do-mcp auth

This will:

1. Ask you to choose Remote (recommended) or Local mode
2. Open your browser to get a Fizzy Personal Access Token
3. Optionally configure an API key for higher rate limits (remote mode)
4. Auto-configure detected AI agents (Claude Desktop, Cursor, OpenCode)

Supported AI Agents

Agent	Status	Local Mode	Remote Mode
Claude Desktop	Supported	Setup Guide	Setup Guide
Cursor	Supported	Setup Guide	Setup Guide
OpenCode	Supported	Setup Guide	Setup Guide

Example Usage

Once configured, your AI assistant can interact with Fizzy naturally:

“Create a new card on the Engineering board titled ‘Implement user authentication’”

“Show me all cards assigned to me that are marked as golden”

“Move card #42 to the ‘In Progress’ column”

“Add a comment to card #123 with my status update”

Architecture

Remote Mode (Hosted Service)

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   AI Agent      │────▶│  Fizzy MCP      │────▶│   Fizzy API     │
│ (Claude, etc.)  │◀────│  Hosted Service │◀────│ (app.fizzy.do)  │
└─────────────────┘     └─────────────────┘     └─────────────────┘
        │                       │
        │                       │
        ▼                       ▼
   MCP over HTTP          Cloudflare Workers
   (Streamable)           (D1 + KV)

Local Mode (Stdio)

┌─────────────────┐     ┌─────────────────┐     ┌─────────────────┐
│   AI Agent      │────▶│   Fizzy MCP     │────▶│   Fizzy API     │
│ (Claude, etc.)  │◀────│  Local Server   │◀────│ (app.fizzy.do)  │
└─────────────────┘     └─────────────────┘     └─────────────────┘
        │                       │
        │                       │
        ▼                       ▼
   MCP Protocol          TypeScript Client
   (JSON-RPC stdio)      (Result types)

Hosted Service

The hosted service at https://fizzy-mcp-hosted.ryanyogan.workers.dev provides:

• Remote MCP Access - Use Fizzy tools from any MCP client via HTTP
• Rate Limiting - Tiered limits based on authentication
• Usage Tracking - Monitor your API usage
• Zero Setup - No local server required

Learn more about the hosted service →

Next Steps

Installation

Learn how to install and configure Fizzy MCP. Read the guide →

Hosted Service

Explore the hosted MCP service and API. Learn more →

Tools Reference

Explore all 40 available MCP tools. Browse tools →

REST API

Access usage stats and admin endpoints. View API docs →