AI web app · Publishing
StoryArc: an AI that interviews you, then helps turn your answers into a book
A web app for people who have a story but not a manuscript. An AI interview draws the story out, then the app helps write it, design a cover and prepare files ready for Amazon KDP. Public launch planned for late 2026.
Visit site
01
An AI interview that turns your answers into a book outline
02
Four steps from idea to KDP-ready PDF, EPUB and DOCX
03
Japanese and English; launch planned for late 2026
01
The problem
Plenty of people have expertise or a life story worth publishing but never get past the blank page, and the route from draft to cover to a publishable file is spread across many tools.
02
How we approached it
- 1
Made the interview the front door: you answer questions and the AI builds the structure of your book from your own words.
- 2
Offered three ways to write (fully AI, AI-assisted or manual) inside one four-step flow: choose an approach, write and edit, design and refine, then publish or print.
- 3
Split each AI job (planning, chapters, covers, blurbs, mockups) into its own server function, so one slow step never blocks the rest.
- 4
Built a switchable AI provider layer with usage tracking and rate limits, so models can change without rewriting the product.
- 5
Added a Japanese/English interface, PDF/EPUB/DOCX export and subscription billing.
03
Challenges we faced
Streaming AI output reliably
Book outlines stream in as they are written. We reworked the parsing and fixed "stuck generating" states so the screen always reflects what the AI is actually doing.
Image generation that fails gracefully
Cover and mockup generation falls back cleanly when an image model errors or hits a limit, instead of breaking the flow.
Interviewing on a phone
The interview was redesigned to run full-screen and responsive on mobile, where most people start.
Project gallery
Have a similar challenge?
Start with a free consultation — we reply within hours.
Get a free consultation →More work
E-commerce · Bilingual Shopify store
AOTOMORI: a bilingual sandal store where customers configure every pair
- 7 models and ~1,090 combinations, each showing only its own photos
- Japanese and English store with sizes in centimetres
- Store build, LINE setup, launch assets and ongoing support from one team
D2C service · AI + e-commerce
FIGURO: from one photo to a custom figure, with a preview before you pay
- An AI preview in four styles in about 30 seconds, before you pay
- One system from photo upload to 3D printing and delivery
- Customer photos deleted automatically 14 days after shipping