2026-07-09 –, Hackers Lab
In 90 minutes, every participant walks out with a runnable local-first app on their own machine: a document model and editor they designed, built, and reviewed using an AI-assisted pipeline, with the option to share peer-to-peer or deploy to the cloud environment (Vetra.)
What participants build
A working document model + editor combo running locally. The app uses Powerhouse document models as the foundational primitive: the schema and the operation log are the application. AI agents act inside the operation log as typed-operation dispatchers.
Participants bring their own idea. They leave with:
- Required: a working local drive on their machine with their document model loaded, their editor running, and at least one document instance dispatched through the full operation log.
- Optional: a GitHub repo of the project they can share peer-to-peer.
- Stretch: a Vetra deployment URL so others can use it without installing anything.
The Powerhouse pipeline (workshop spine)
Participants run their idea through 4 stages, with agent support each step of the way:
- Functional analysis
- Data architecture
- Source code / build
- QA checks & published package
Every stage has an explicit input, explicit output, and a handoff phrase to the next agent. The workflow is guided by the workshop facilitator. This is the Powerhouse BA Praxis methodology applied to a participant's own project, in real time.
Wouter has worked on Defi at MakerDAO since 2017, before shifting focus to decentralized organizations and local-first application building at Powerhouse.