Trio can connect to AI assistants - such as Claude - over the Model Context Protocol (MCP), letting you ask an assistant questions about your studio in plain language. This article explains what you need to turn it on, how the connection works, exactly what the assistant can see, and how to review or revoke access at any time.
Connecting an AI assistant is a studio-owner feature with a few requirements:
Trio exposes a secure MCP endpoint. When you connect an AI assistant, you authorize it through Trio's sign-in flow - the same login you already use. No password or API key is ever shared with the assistant.
The connection is read-only. A connected assistant can look up studio information to answer your questions - schedules, students and families, invoices, and similar records scoped to your studio and your role. It uses the same multi-tenant boundaries as the rest of Trio, so an assistant connected to your studio can never read another studio's data.
The assistant cannot create, edit, or delete anything in Trio. It cannot send messages, charge cards, change schedules, or modify records. If you want to make a change, you do it yourself in Trio.
Every connected assistant appears under Settings → Connected apps. From there an owner or admin can see which assistants are connected, when each was last active, and revoke any connection.
Revoking a connection takes effect immediately - the assistant loses access the next time it tries to read your studio data. You can reconnect later by authorizing again.
Once data leaves Trio and reaches a connected AI assistant, it is handled by that assistant's provider under their terms and privacy policy. Trio cannot controlhow a third-party assistant stores, logs, or uses the information it receives. Only connect assistants you trust, and revoke access from Settings → Connected apps if you no longer use one.