About Stateless Hono MCP Server
A Hono-based example MCP server that uses Streamable HTTP transport, adapted from the official Express example. It runs statelessly and can be deployed to Cloudflare Workers or any other runtime Hono supports, making it a reference implementation.
Key features
- 01MCP server built with the Hono framework
- 02Streamable HTTP transport
- 03Stateless operation mode
- 04Request conversion via fetch-to-node
- 05Deployable to Cloudflare Workers
Use cases
- 01Learning how to build an MCP server with Hono
- 02Deploying an MCP server to Cloudflare Workers for testing
- 03Checking tools and prompts with the example client
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEStateless Hono MCP Server
An example Hono MCP server using Streamable HTTP, based off the official Express example, using fetch-to-node to convert, deployable to Cloudflare Workers (and anywhere else Hono runs).
The only real changes to the Express example are:
- import express, { Request, Response } from 'express';
+ import { Hono } from 'hono';
+ import { toFetchResponse, toReqRes } from 'fetch-to-node';
// ...
- const app = express();
- app.use(express.json());
+ const app = new Hono();
- app.post('/mcp', async (req: Request, res: Response) => {
+ app.post('/mcp', async (c) => {
+ const { req, res } = toReqRes(c.req.raw);
const server = getServer();
try {
const transport: StreamableHTTPServerTransport = new StreamableHTTPServerTransport({
sessionIdGenerator: undefined,
});
await server.connect(transport);
- await transport.handleRequest(req, res, req.body);
+ await transport.handleRequest(req, res, await c.req.json());
res.on('close', () => {
console.log('Request closed');
transport.close();
server.close();
});
+ return toFetchResponse(res);
} catch (error) {
Testing with an example MCP client
In one terminal:
npm start
In another:
node node_modules/@modelcontextprotocol/sdk/dist/esm/examples/client/simpleStreamableHttp.js
This will try to connect to the MCP server running on port 3000. You can use connect <url>/mcp to connect to a different host or port.
Then you can run commands like list-prompts or list-tools to verify your MCP server is working.
Deploying
npm run deploy
Here's an example session against a Hono MCP server deployed on Cloudflare Workers:
> connect https://mcp-hono-stateless.michael.workers.dev/mcp
Connecting to https://mcp-hono-stateless.michael.workers.dev/mcp...
Transport created with session ID: undefined
Connected to MCP server
> list-tools
Available tools:
- start-notification-stream: Starts sending periodic notifications for testing resumability
> list-prompts
Available prompts:
- greeting-template: A simple greeting prompt template
> call-tool start-notification-stream
Calling tool 'start-notification-stream' with args: {}
Notification #1: info - Periodic notification #1 at 2025-04-22T16:20:50.178Z
>
Notification #2: info - Periodic notification #2 at 2025-04-22T16:20:50.278Z
>
Notification #3: info - Periodic notification #3 at 2025-04-22T16:20:50.378Z
>
Notification #4: info - Periodic notification #4 at 2025-04-22T16:20:50.478Z
>
Notification #5: info - Periodic notification #5 at 2025-04-22T16:20:50.578Z
>
Notification #6: info - Periodic notification #6 at 2025-04-22T16:20:50.678Z
>
Notification #7: info - Periodic notification #7 at 2025-04-22T16:20:50.778Z
>
Notification #8: info - Periodic notification #8 at 2025-04-22T16:20:50.878Z
>
Notification #9: info - Periodic notification #9 at 2025-04-22T16:20:50.978Z
>
Notification #10: info - Periodic notification #10 at 2025-04-22T16:20:51.078Z
> Tool result:
Started sending periodic notifications every 100ms
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.