Skip to Content
Connect your AI assistant

Connect your AI assistant

You can connect your own AI assistant — Claude, ChatGPT, Cursor, and others — to your PlanMyPeak account and ask it questions about your coaching in plain language: “Who missed a workout this week?”, “Summarise Alex’s last month.” The assistant reads your data live and answers, so you don’t have to click through screens to pull the picture together.

Read-only, and only your athletes. A connected assistant can look at your coaching data — it can never change anything, send anything to your athletes, or see athletes you don’t coach. You approve the connection once with your own PlanMyPeak sign-in; no passwords or keys are ever shared with the assistant.

What you can ask about

Once connected, your assistant can look up:

  • Your athletes — who’s on your roster, with name/email search and status.
  • Workouts — planned and completed sessions for an athlete over any date range.
  • A single activity — the full detail of one workout: planned vs. actual, power, heart rate, and the analysis.
  • The training plan — what’s scheduled for an athlete over a window.
  • Weekly compliance — planned vs. completed load and compliance over recent weeks.

Your units carry across. Distances, elevation, and temperatures in assistant answers use the measurement units saved in your PlanMyPeak settings. A coach and athlete can ask about the same workout and each receive values in their own selected units.

Before you start

  • The connection address (the “MCP server URL”) is: https://api.planmypeak.com/mcp
  • You’ll sign in with your normal PlanMyPeak account and approve read-only access. The assistant receives a secure connection that renews itself — there’s nothing to copy, paste, or refresh.
  • You can disconnect it at any time from your assistant’s own settings.

Some assistants need a paid plan. Adding a custom connector is a paid-tier feature in several apps (for example, custom connectors on Claude are available on Free through Enterprise, with Free limited to one; ChatGPT requires a paid plan with developer mode). This is a limit of the assistant app, not of PlanMyPeak. Custom connectors are a newer, beta feature in some apps — menus may move around a little.

Connect it to your assistant

The idea is the same everywhere: add a custom connector (also called an “MCP server” or “app”) and give it the address above. Here are the popular ones.

Claude (web and desktop)

  1. Open Settings → Connectors.
  2. Select Add custom connector.
  3. Paste https://api.planmypeak.com/mcp as the remote MCP server URL and select Add.
  4. Back in a chat, enable the PlanMyPeak connector, then follow the sign-in prompt (see Signing in below).

On a Team or Enterprise workspace, an Owner adds the connector once for the workspace; each coach then connects to it individually and signs in with their own PlanMyPeak account — so everyone only ever sees their own athletes.

Claude Code (command line)

Run:

claude mcp add --transport http planmypeak-coach https://api.planmypeak.com/mcp

Then start a session (or run /mcp) and connect — your browser opens for sign-in.

ChatGPT

Custom MCP connectors live behind developer mode (a paid plan is required):

  1. In Settings, open the connectors/apps area and turn on developer mode.
  2. Add a new connector/app and enter https://api.planmypeak.com/mcp as the MCP server URL.
  3. Complete the sign-in prompt to authorise read-only access.

Cursor

  1. Open Settings → Tools & MCP → New MCP Server.

  2. Add a Streamable-HTTP server pointing at https://api.planmypeak.com/mcp, for example:

    { "mcpServers": { "planmypeak-coach": { "url": "https://api.planmypeak.com/mcp" } } }
  3. Enable it; Cursor will prompt you to sign in.

Other assistants

Any tool that supports remote MCP servers over Streamable HTTP works the same way: add a custom server with the address https://api.planmypeak.com/mcp. The tool handles the secure sign-in for you.

Signing in and approving

The first time you connect, your assistant sends you to PlanMyPeak to authorise:

  1. A browser window opens and asks you to sign in to PlanMyPeak (if you’re not already signed in) — the same login you use for the app.
  2. You’ll see an approval screen naming the assistant and stating that it’s asking for read-only access to your coaching data.
  3. Select Allow access. You’re returned to your assistant, now connected.

That’s it — no tokens to paste, and the connection renews itself so you won’t have to repeat this.

Examples of real use

Talk to your assistant the way you’d talk to an assistant coach. Use athlete names and dates naturally — it maps them to your roster.

  • “Which of my athletes missed a planned workout this week?”
  • “Give me a compliance summary for Alex over the last 4 weeks — planned vs. completed load — and flag any week under 80%.”
  • “Show me the details of Sam’s ride yesterday: normalised power, heart rate, and how it compared to the plan.”
  • “What’s on Maria’s plan for the next 7 days?”
  • “Across my whole roster, who had the biggest jump in training load this week?”
  • “List everyone who’s behind on compliance and draft a short, encouraging check-in message I could send each of them.”
  • “Who on my roster hasn’t completed a workout in the last 10 days?”

Because the assistant reads live data, the answers reflect the latest workouts as they sync in.

The assistant only reads — you still act. It can pull together a check-in message or a summary, but it never sends anything to your athletes or changes a plan. You decide what to do with what it finds, inside PlanMyPeak or your TrainingPeaks workflow.

Privacy and control

  • Read-only. A connected assistant can never modify your data, publish feedback, or act on your behalf.
  • Scoped to you. It sees only the athletes you coach — exactly what you see in PlanMyPeak — and nothing from other coaches.
  • You approve each connection with your own sign-in, and you can disconnect it whenever you like from your assistant’s connector settings.
  • Your PlanMyPeak password is never shared with the assistant.

Have a question that isn’t covered here? See the FAQ.

Last updated on