Authorization Server vs. Resource Server

These are two jobs in the login flow your connector uses inside ChatGPT. The authorization server is the bouncer: it checks who the buyer is and hands back a pass. The resource server is the thing that actually does the work — fetching data, booking the meeting — but only after it sees a valid pass. Splitting them means the part that proves identity stays separate from the part that holds your data.

Why it matters for the ChatGPT funnel

This is mostly under the hood, but here is the one thing you need to know. When these two roles are wired up wrong, the buyer gets bounced to a broken login or a permission error right as they try to book. That is a lost lead at the exact moment of intent. Clean separation keeps the handoff invisible, so the chat flows from interest to booked meeting without a stumble.

How drio fits

You do not have to think about this. drio runs both roles for your connector, so the mcp-server that books your leads only ever talks to verified buyers. You watch pipeline land; we keep the login plumbing quiet.