From "hello" to a
live app on your domain

The macro flow

Four stages, one continuous path

Stage 1

Onboarding — IQ learns the expert

Before anything is built, the platform builds a model of the expert: who they are, what they sell, and what they know. This is what makes the generated outputs actually good.

1. Session & dossierExpert starts a session; IQ captures their background, story, and positioning into a structured dossier.
2. Offer stackThe agent generates the business offer and pricing tiers — editable, section by section.
3. Context profile & avatarsICP and market positioning are modeled; FAL generates a portrait per avatar.
4. Knowledge baseFiles, URLs, and pasted text are extracted, embedded, and indexed — with a manual review checkpoint.
5. QA checkpoint → handoffAn LLM validation pass; the expert approves; a default workspace is created and the engines are selected.

Stage 2 & 3

Build & preview — the agent does the work

The expert asks IQ for something — "build me a treatment cost estimator." IQ picks the right tools, spins up a sandbox, writes the app, and returns one clean preview URL.

Stage 4

Publish & deploy — one approval, live

When the expert is happy, publish freezes the build into an immutable release and deploy promotes it to a stable production URL on cliniciq.com. Production never changes until they republish.

The data plane, in parallel

When an app needs a database

For full-stack apps, the same flow provisions managed data and wires it in automatically — proven in the end-to-end test on the Infrastructure page.