Not another subscription, not your data inside someone else's model. A small group that meets live every week, builds together, and stays a community long after the 8 weeks. You ship your own project with peers and walk away owning the system.
€300 founding price · 8 weeks live · A small community you keep
Solo founders, fractional execs, and operators who already use AI every day and want to own the system, not rent it. You build your own project, on your own files, alongside a small community building the same way, and you stay part of it after the cohort ends.
Not for beginners who have never touched Obsidian or Claude. Not for people who want a finished SaaS to log into. This is build-with-me, not a passive course.
Live Tuesdays, 2 hours each · 17:00 Riga / 16:00 CEST / 15:00 London (BST) / 10:00 New York · 9 June to 28 July 2026 · Recorded if you miss one.
Map how you actually work and set up the knowledge base your system will run on. The first principle: your data stays yours.
Claude and ChatGPT as real thinking partners. The prompting patterns that hold up under pressure, not party tricks.
Turn your own notes and files into something a model can read and reason over. Retrieval grounded in your work, not the public internet.
Connect the pieces and ship your first workflow that runs without you in the loop.
Put a real front end on your system so it is something you use daily, not a pile of scripts.
Edit and extend your system directly. Still no heavy coding, just enough to make it yours.
Finish the concrete thing you came to build, with the group around you and feedback in real time.
The patterns to extend your system for whatever model comes next, and a clear path for where you go from here.
First small group, special founding price. Future cohorts at normal price.
€300one time
Founding-cohort members get year-one platform access included.
Members of the founding cohort get first access to future events and pricing.
Only 6 seats left
Join the founding cohortWhat's said in the cohort stays in the cohort. We transcribe calls for participants' own learning, never publish without consent.