Solutions — Playbooks
Never re-explain
a process again.
A playbook is a saved, step-by-step process — a deploy checklist, a client onboarding flow, anything you’ve done once. ChatGPT or Claude runs it with you from step one, every time, so it never drifts or half-remembers a past run.
Stateless by design
Every run starts at step one
Playbooks never resume where a past conversation left off — unlike the curriculum, which tracks your progress across sessions, a playbook always runs fresh. Run it once a week or ten times a day; it behaves the same every time.
One step at a time
Instructions, then a check
Your AI presents one step with its instructions and an optional “done when” check, records it, and moves to the next when you say you’re done. The check is informative only — it’s never a gate you have to pass.
What Running One Looks Like
Say the word, your AI takes over
You
Run one of my MCP Tutor playbooks.
ChatGPT or Claude
You have three: “Ship a React site to Vercel,” “New client onboarding,” and “Weekly report draft.” Which one?
You
Ship a React site to Vercel.
ChatGPT or Claude
Step 1 of 6: Confirm the build passes locally with npm run build. Done when: no errors in the output. Let me know when you’re there.
Build your own
Save what you just did
Write one on the web dashboard, or just ask your AI to save the process you two just walked through as a new playbook — it names the steps and captures what each one needs.
Ready-made
Example templates included
Start from a library of built-in example playbooks — read-only until you use one, at which point you get your own editable copy to adjust.
Shared with your team
One person builds it, everyone runs it
Share a playbook org-wide so a process one person figured out becomes something the whole team can run the same way. See MCP Tutor for Teams.