Workers + Stytch TODO App MCP Server

This is a Workers server that composes three functions:

A static website built using React and Vite on top of Worker Assets
A REST API built using Hono on top of Workers KV
A Model Context Protocol Server built using on top of Workers Durable Objects

User and client identity is managed using Stytch. Put together, these three features show how to extend a traditional full-stack application for use by an AI agent.

This demo uses the Stytch Consumer product, which is purpose-built for Consumer SaaS authentication requirements. If you are more interested in Stytch's B2B product, see this demo instead.

Set up

Follow the steps below to get this application fully functional and running using your own Stytch credentials.

In the Stytch Dashboard

Create a Stytch account. Within the sign up flow select Consumer Authentication as the authentication type you are interested in. Once your account is set up a Project called "My first project" will be automatically created for you.
Navigate to Frontend SDKs to enable the Frontend SDK in Test
Navigate to Connected Apps to enable Dynamic Client Registration and configure your authorization URL as http://localhost:3000/oauth/authorize.
Navigate to Project Settings to view your Project Domain, Project ID, and API keys. You will need these values later.

On your machine

In your terminal clone the project and install dependencies:

git clone https://github.com/cloudflare/ai.git cd ai npm i cd demos/mcp-stytch-consumer-todo-list

Next, create an .env.local file by running the command below which copies the contents of .env.template.

cp .env.template .env.local

Open .env.local in the text editor of your choice, and set the environment variables using the public_token and project domain found on Project Settings.

# This is what a completed .env.local file will look like VITE_STYTCH_PUBLIC_TOKEN=public-token-test-abc123-abcde-1234-0987-0000-abcd1234 VITE_STYTCH_DOMAIN=random-domain-name.customers.stytch.dev

Create a .dev.vars file by running the command below which copies the contents of .dev.vars.template

cp .dev.vars.template .dev.vars

Open .dev.vars in the text editor of your choice, and set the environment variables using the Project ID and Secret found on Project Settings.

// This is what a completed .dev.vars file will look like STYTCH_PROJECT_ID=project-test-6c20cd16-73d5-44f7-852c-9a7e7b2ccf62

Related MCP server: Remote MCP Server (Authless)

Running locally

After completing all the setup steps above the application can be run with the command:

npm run dev

The application will be available at http://localhost:3000 and the MCP server will be available at http://localhost:3000/sse.

Test your MCP server using the MCP Inspector

npx @modelcontextprotocol/inspector@latest

Deploy to Cloudflare Workers

Click the button - you'll need to configure environment variables after the initial deployment.

Deploy to Cloudflare

Or, if you want to follow the steps by hand:

Create a KV namespace for the TODO app to use

wrangler kv namespace create TODOS

Update the KV namespace ID in wrangler.jsonc with the ID you received:

"kv_namespaces": [ { "binding": "TODOS", "id": "your-kv-namespace-id" } ]

Upload your Stytch Env Vars for use by the worker

npx wrangler secret bulk .dev.vars

Deploy the worker

npm run deploy

Grant your deployment access to your Stytch project. Assuming your Stytch project was deployed at https://mcp-stytch-consumer-todo-list.$YOUR_ACCOUNT_NAME.workers.dev:
1. Add https://mcp-stytch-consumer-todo-list.$YOUR_ACCOUNT_NAME.workers.dev/authenticate as an allowed Redirect URL
2. Add https://mcp-stytch-consumer-todo-list.$YOUR_ACCOUNT_NAME.workers.dev as an allowed Authorized Application in the Frontend SDKs configuration