Forge Agents
← OKA platform
Coming soon

MOKA — grounded answers for
medical front desks, without ever playing doctor.

The OKA engine, tuned for medical offices. MOKA answers your staff's "how do we do it here?" questions only from your practice's own documented procedures — scheduling rules, intake workflows, office policies — and cites the exact source, every time. It is staff-facing, never patient-facing, and it does not answer clinical questions. Not hedged, not "probably" — it refuses and routes them to your licensed staff.

Request early access
Try it — scripted preview

Ask MOKA a front-desk question

A scripted sample on a fictional medical office's procedures (“Lakeside Family Medicine”). Pick a question — and watch it refuse where it should. Your deployment answers live, from your own documents.

MOKA · Lakeside Family Medicine (demo corpus)scripted preview
Pick a question above to see MOKA answer from the demo procedures.

Scripted demo content. In your deployment, MOKA answers from your documents — with every citation machine-verified.

Why MOKA

Medical is where the refusal matters most.

A front desk that guesses is a liability anywhere. In a medical office it is unacceptable — so MOKA is built so the model can only speak from what your practice has actually documented, and clinical territory is walled off entirely.

📎

Cited to your source

Every answer points to the internal procedure or policy it came from — so it's verifiable, not trusted blindly.

🚫

Never clinical, never triage

Symptom, urgency, and treatment questions are refused outright and routed to your licensed staff. That boundary is built into MOKA, not configured onto it.

🔒

Confidential by design

MOKA reads your procedures, not patient charts. Your data is never used to train shared models; minimal collection, locked-down access.

Be first

MOKA is in development.

Medical is the vertical we are building most carefully, and it ships last on purpose. Want to help shape it? Tell us about your office and we'll bring you into the early cohort.

Request early access