Voice + Text Interview Engine
OpenAI Whisper + GPT-4 build a locked BookSpec JSON from a natural conversation.
A full-stack AI book creation and print-on-demand SaaS for a US creative-tech startup — users author books through voice or text interviews, generate AI illustrations and character art, review a rendered manuscript, and complete purchase, all the way to physical hardcover delivery via the Lulu Print API.
What the work moved — the proof points that came out of the build.
Aspiring authors faced a fragmented, high-friction journey — writing tools, cover design, formatting and print fulfilment all lived in disconnected products. There was no single platform that could take a raw creative idea from first word to physical book without requiring design or publishing expertise.
We scoped a five-frame experience that hid the complexity of generative AI behind a single linear journey: interview, generate, review, pay, fulfil. Each frame had to deliver a complete artefact the author could approve before the next began.
The platform pairs an OpenAI Whisper + GPT-4 interview engine — building a locked BookSpec JSON from voice or text — with a real-time manuscript and artifact generator producing character sketches and environment art via DALL-E. A DocRaptor-powered renderer generates 2-page spread previews; Stripe handles billing and subscriptions; and a live 4-stage fulfilment tracker is driven by Lulu Print API webhooks. Everything sits on a Python / PostgreSQL API with AWS S3 media storage.
OpenAI Whisper + GPT-4 build a locked BookSpec JSON from a natural conversation.
Character sketches and environment art generated in real time via DALL-E.
DocRaptor-powered renderer produces 2-page spread previews before purchase.
End-to-end checkout and recurring billing inside the same flow.
Live status driven by Lulu Print API webhooks from order through delivery.
Tell us what's slowing your team down. We'll come back with a scoped, no-obligation plan.