A paid 30-day sprint with the founder: pick one recurring AI-assisted workflow and turn it into versioned capability the factory can project into the runtimes you already use. This is a design-partner path, not a self-serve product. Scaffolde is still private, single-operator, macOS.
The public factory story stays Author, Validate, Project, Evidence. The trial is how a team uses that loop on one real workflow.
Pick one recurring workflow tied to revenue, risk, or senior time.
Name the owner, inputs, approvals, success metric, and data boundary.
Author it as canonical source, then project it into the surfaces you already run.
Four weekly reviews. Visible state, human review, evidence back into source.
Expand, hold, or stop. The pilot fee credits toward an annual plan if both sides continue.
Someone who can judge whether the workflow actually matters.
One operator who knows the current process and can review outputs.
Real artifacts or safe sample data. Sensitive data is not required to scope.
Cycle time, expert time saved, accepted outputs, risk reduction, or throughput.
Four short reviews to inspect runs, failure modes, and value evidence.
What can be used, what stays out, and what needs approval.
Scoped to one workflow. The pilot fee can be credited toward an annual plan if both sides choose to continue. Pricing is set after workflow fit, data boundary, stakeholder load, and integration needs are clear.
No form backend yet. Request a trial by email, or book a call.
Walk the factory model on a workflow you already run. Expand only when the evidence says to.