No Local Server
Skip installation and configuration. Connect directly from your AI agent without running anything locally.
The Fizzy MCP Hosted Service provides remote MCP access without requiring you to run a local server. Your AI assistant connects directly to our cloud infrastructure, giving you instant access to all Fizzy MCP tools from anywhere.
The hosted service is a fully managed MCP server running on Cloudflare’s global edge network. Instead of running npx @fizzy-mcp/server locally, your AI agent connects via HTTP to our hosted endpoint:
https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcpThis endpoint speaks the Model Context Protocol over HTTP Streamable transport, making it compatible with any MCP client that supports HTTP-based connections.
No Local Server
Skip installation and configuration. Connect directly from your AI agent without running anything locally.
Always Updated
Automatically running the latest version. No manual updates or version management required.
Works Anywhere
Access from any device or environment with an internet connection. Perfect for remote development, cloud IDEs, and mobile workflows.
Zero Maintenance
We handle uptime, scaling, and infrastructure. Focus on using Fizzy, not managing servers.
The hosted service is built on a modern, globally distributed architecture designed for reliability and low latency:
┌─────────────────────────────────────────────────────────────────┐│ Your AI Assistant ││ (Claude, Cursor, OpenCode) │└─────────────────────────┬───────────────────────────────────────┘ │ │ MCP over HTTP │ (Streamable Transport) ▼┌─────────────────────────────────────────────────────────────────┐│ Cloudflare Workers Edge ││ ││ ┌───────────────┐ ┌───────────────┐ ┌───────────────┐ ││ │ Worker │ │ Worker │ │ Worker │ ││ │ (US-West) │ │ (EU-West) │ │ (Asia-East) │ ││ └───────────────┘ └───────────────┘ └───────────────┘ ││ ││ ┌───────────────────────────────────────────────────────────┐ ││ │ D1 Database │ ││ │ (Rate limits, usage tracking, analytics) │ ││ └───────────────────────────────────────────────────────────┘ ││ ││ ┌───────────────────────────────────────────────────────────┐ ││ │ KV Storage │ ││ │ (API keys, session data, caching) │ ││ └───────────────────────────────────────────────────────────┘ │└─────────────────────────┬───────────────────────────────────────┘ │ │ HTTPS ▼┌─────────────────────────────────────────────────────────────────┐│ Fizzy API ││ https://app.fizzy.do │└─────────────────────────────────────────────────────────────────┘| Component | Purpose |
|---|---|
| Cloudflare Workers | Serverless compute at the edge, handling MCP protocol translation |
| D1 Database | SQLite-compatible database for rate limiting and usage analytics |
| KV Storage | Key-value store for API key validation and session management |
| Workers AI | Optional AI features for enhanced tool processing |
| Endpoint | Method | Description |
|---|---|---|
/mcp | POST | Main MCP endpoint for tool calls |
/health | GET | Health check endpoint |
https://fizzy-mcp-hosted.ryanyogan.workers.devThe hosted service requires your Fizzy Personal Access Token to authenticate requests to the Fizzy API. This token is passed via the X-Fizzy-Token HTTP header.
| Header | Required | Description |
|---|---|---|
X-Fizzy-Token | Yes | Your Fizzy Personal Access Token (starts with fz_) |
X-API-Key | No | Optional API key for higher rate limits (starts with fmcp_) |
Content-Type | Yes | Must be application/json |
You configure your AI agent with your Fizzy token in the headers
Your AI agent sends an MCP request to the hosted service
The hosted service validates the request format and checks rate limits
The hosted service forwards the request to the Fizzy API using your token
The Fizzy API returns the response, which is relayed back to your AI agent
Get up and running with the hosted service in minutes:
Get your Fizzy token
Go to app.fizzy.do, navigate to Profile → API, and create a Personal Access Token.
Test the connection
Use curl to verify your token works:
curl -X POST https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcp \ -H "Content-Type: application/json" \ -H "X-Fizzy-Token: fz_your_token_here" \ -d '{ "jsonrpc": "2.0", "id": 1, "method": "tools/list", "params": {} }'You should receive a JSON response listing all available tools.
Configure your AI agent (see examples below)
Start using Fizzy
Open your AI agent and try a command like:
“List all my Fizzy boards”
Add to your Claude Desktop config (~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{ "mcpServers": { "fizzy": { "transport": "http", "url": "https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcp", "headers": { "X-Fizzy-Token": "fz_your_token_here" } } }}Add to your Cursor MCP settings (~/.cursor/mcp.json):
{ "mcpServers": { "fizzy": { "transport": "http", "url": "https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcp", "headers": { "X-Fizzy-Token": "fz_your_token_here" } } }}Add to your OpenCode config (~/.config/opencode/opencode.json):
{ "mcp": { "fizzy": { "type": "http", "url": "https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcp", "headers": { "X-Fizzy-Token": "fz_your_token_here" } } }}Here’s a complete example of calling the fizzy_list_boards tool via the hosted service:
curl -X POST https://fizzy-mcp-hosted.ryanyogan.workers.dev/mcp \ -H "Content-Type: application/json" \ -H "X-Fizzy-Token: fz_your_token_here" \ -d '{ "jsonrpc": "2.0", "id": 1, "method": "tools/call", "params": { "name": "fizzy_list_boards", "arguments": {} } }'{ "jsonrpc": "2.0", "id": 1, "result": { "content": [ { "type": "text", "text": "[{\"id\":\"abc123\",\"name\":\"Product Roadmap\",\"card_count\":42}]" } ] }}The hosted service implements rate limiting to ensure fair usage. See the Rate Limits page for detailed information.
Quick overview:
| Tier | Writes/Day | Reads/Day |
|---|---|---|
| Anonymous | 100 | 1,000 |
| Authenticated (API Key) | 1,000 | 10,000 |
| Admin | Unlimited | Unlimited |
| Feature | Hosted Service | Local Server |
|---|---|---|
| Setup | None (just configure headers) | Install Node.js, run npx |
| Updates | Automatic | Manual (npm update) |
| Rate Limits | Yes (100-1,000 writes/day) | None |
| Offline Access | No | Yes (after auth) |
| Token Storage | Sent with each request | Stored locally |
| Latency | Network roundtrip | Direct stdio |
| Best For | Remote dev, cloud IDEs, simplicity | High-volume, offline, privacy |
X-Fizzy-Token header is set correctlyfz_X-RateLimit-Reset header for when limits reset*.workers.dev