Skip to main content

​
Overview

ChatGPT Apps allow you to extend ChatGPT with custom functionality by connecting to external services. Mirobody Health supports ChatGPT Apps through its MCP endpoint, giving ChatGPT access to health data retrieval, provider connections, and more.
Learn more about ChatGPT Apps in the official announcement.

​
Prerequisites

Before connecting Mirobody Health to ChatGPT, you need to make your local MCP endpoint publicly accessible:
1

Ensure Mirobody Health is running

Your Mirobody Health instance should be running and accessible at http://localhost:18080/mcp.
2

Expose your local endpoint

Use a tunneling service to make your local endpoint publicly accessible:
# Using ngrok
ngrok http 18080

# Or using cloudflared
cloudflared tunnel --url http://localhost:18080
This will provide you with a public URL (e.g., https://abc123.ngrok.io) that forwards to your local instance.
3

Note your public MCP URL

Your public MCP endpoint will be: https://your-tunnel-domain/mcp
Keep this URL secure. Anyone with access to this URL can interact with your Mirobody Health instance.

​
Creating a ChatGPT App

Once your MCP endpoint is publicly accessible, you can create a ChatGPT App:
1

Access ChatGPT Apps

Open ChatGPT and navigate to the Apps section. You’ll see your existing apps and the option to create new ones.
ChatGPT Apps main interface showing available apps
2

Enable Developer Mode

Click on “Create your own” or access Developer Mode to start building a custom app. This will open the app creation interface where you can configure your Mirobody Health connection.
ChatGPT Developer Mode for creating custom apps
3

Add your MCP endpoint

In the app configuration:
  1. Name your app: Give it a descriptive name like “Mirobody Health”
  2. Add MCP endpoint: Enter your public MCP URL: https://your-tunnel-domain/mcp
  3. Description (optional): Add a description of what health capabilities this app provides
  4. Test connection: ChatGPT will validate that it can connect to your MCP endpoint
Adding a new ChatGPT app with MCP endpoint configuration
ChatGPT will automatically discover all available tools from your Mirobody Health instance once connected.
4

Authenticate via OAuth

When ChatGPT first connects to your Mirobody Health app, it will automatically redirect you to the login page for OAuth authentication.
Mirobody Health has built-in OAuth support. Simply log in with your credentials when prompted, and the authentication flow will complete automatically.
This OAuth flow ensures secure access to your health data and only needs to be completed once per ChatGPT session.
5

Start using your app

Once authenticated, your ChatGPT session can now:
  • Retrieve user health profiles
  • Query health data from connected providers
  • Access any custom tools you’ve added to Mirobody Health
  • Use resources exposed through your MCP server
Your ChatGPT App is now connected to Mirobody Health!

​
Available Capabilities

When connected, ChatGPT will have access to:

Health Profile

Retrieve comprehensive user health data and connected device information

Health Data Query

Query specific health indicators and time-series data

Custom Tools

Any custom tools you’ve added to your Mirobody Health instance

MCP Resources

Access to resources exposed through your MCP implementation

​
Example Use Cases

Once connected, ChatGPT can perform sophisticated health data analysis by pulling information from all your connected devices, health records, and other relevant sources. Here are some real-world examples:

​
Quick Queries

For simple questions, you can ask: 
"Show me my health profile"
"What was my heart rate yesterday?"
"Compare my sleep data from this week to last week"
"What health devices do I have connected?"

​
Advanced Analysis: Diabetes Management

Prompt: “Summarize the history and recent progress of my diabetes. Use Theta to find data from devices, records, and other relevant sources.”

View Full Conversation

See the complete diabetes analysis conversation

Interactive Chart

Explore the interactive glucose trend visualization

​
Advanced Analysis: Pain Pattern Recognition

Prompt: “Help me understand the whole picture and the progress of my knee pain. Use all relevant data and find potential drivers. Use visual illustrations if it’s helpful.”

View Full Analysis

See how ChatGPT analyzes complex pain patterns with multi-source data
These examples demonstrate ChatGPT’s ability to synthesize data from multiple sources, perform longitudinal analysis, and generate actionable insights - all powered by your Mirobody Health data.

​
Security Considerations

OAuth Authentication Built-in:Mirobody Health includes OAuth authentication, which means users must log in before ChatGPT can access their health data. This provides a secure authentication layer for all requests.
Additional Security Recommendations:
  • Your public MCP endpoint URL should be kept secure and only shared with trusted services
  • Consider implementing rate limiting at the tunnel level
  • Monitor access logs for unusual activity
  • Only expose your endpoint when actively using it with ChatGPT
  • For production use, consider additional security measures like IP whitelisting
  • OAuth tokens have expiration times - users will need to re-authenticate when tokens expire

​
Troubleshooting

​
Connection Failed

If ChatGPT cannot connect to your MCP endpoint:
  1. Verify your tunnel is running and accessible
  2. Check that Mirobody Health is running on port 18080
  3. Test the MCP endpoint directly: curl https://your-tunnel-domain/mcp
  4. Ensure no firewall is blocking the connection

​
Tools Not Appearing

If tools aren’t showing up in ChatGPT:
  1. Verify your MCP endpoint returns tool definitions
  2. Check Mirobody Health logs for errors
  3. Restart your tunnel and try reconnecting
  4. Ensure your tools are properly registered in Mirobody Health

​
Next Steps

Built-in Tools

Explore available health tools

Adding Custom Tools

Create your own custom tools

MCP Protocol

Learn about the MCP protocol

Security Best Practices

Production deployment guide