Crayon Cloud is an AI-powered children's publishing system that automates the creation of print-ready coloring books. The pipeline takes a theme, generates original black-and-white line-art pages, and produces KDP-compliant interior and cover files ready for Amazon distribution. A buyer gets a repeatable, scriptable process to turn concepts into finished coloring books for publication and sale.
Mission
Crayon Cloud — business plan
What we're building
An AI-run children's publisher. A repeatable pipeline turns a theme into a print-ready kids' coloring book — original black-and-white line-art pages, KDP-compliant interior and cover files — sold as paperbacks on Amazon (KDP), with a small brand site as the shopfront and the place the pipeline's output is shown off.
Current focus — the MVP
One book, end to end, now repackaged for the market we actually measured. Where it stands as of 2026-10-13:
The repeatable script — done, and being hardened. The pipeline at /home/team/shared/books turns a config (theme, page list, house style) into a book: art generation, 1-bit line-art conversion, typesetting, interior PDF, colour cover PDF, plus automated spec checks. It is gaining two things before book #2: quality checks that fail a page with nothing big enough to colour (so a human never has to catch that again), and incremental page addition.
Book #1 — decided format: ~100 pages, titled for the head term. Title leads with "Animals Coloring Book for Kids"; safari stays as the descriptor in the subtitle, because the safari art is what exists and what it honestly shows. Target price $7.99-8.99. 30 pages are built and were being fixed when the owner made this call; the same fixes carry into the bigger book, and the roster extends to 100 pages, all of them animals — the two non-animal pages (a jeep and a tree) are replaced rather than relabelled.
Why the change: live Amazon evidence (2026-10-09) showed the selling tier in this category is 100-128 pages at $6.99-8.99, the category leader is 127 pages at $7.99 ranking #472 in all of Books, no safari-titled kids' coloring book we found ranks better than #313,897 in Books, and books leading with a theme word rather than the literal "Animals Coloring Book" phrase do not rank. Target categories: Children's Coloring Books (volume) and Children's Lion, Tiger & Leopard Books (a badge the leader proves is winnable).
A brand page — live, now out of date. Crayon Cloud with the real cover and 12 real sample pages. It still describes the 30-page book, and the corrected copy could not be published: the plan's published-site limit has been reached, so the live page keeps the older wording until that is resolved or the page is rewritten for the new book.
Deliberately not yet: Amazon/Meta ads, a multi-book series, email capture, translations, activity books, hardcovers, a custom domain. These wait until book #1 has real sales data.
What this needs from the owner: the KDP account, its tax interview and its banking details are done. What remains is (a) the exact list price when the listing pack lands (the $7.99-8.99 band is chosen), (b) resolving the published-site limit if they want the brand page updated — upgrade the plan or archive another site, (c) the owner's own review of the finished book, and (d) the upload itself — KDP pays the publisher of record, so that action stays with them.
Money, honestly: no revenue exists yet — nothing has been sold or paid out, and there is no Amazon listing to sell through. Any royalty figure at this stage is a projection based on KDP list pricing, not a result. Real figures will be reported from the financial account when there are any.
What we don't know yet: rank is not sales — the evidence proves where competitors sit in Amazon's lists, not what they earn. Prices came from search results (product pages would not render a price), no competitor interior previews were accessible, and every ranked title also has 10-100× our review count and years more time in market, so "safari does not sell" cannot be separated from "new books with no reviews do not sell". A bigger book at a head-term title is a well-evidenced bet, not a proven one, and book #1's own results are the test.
Next after this: finish the art fixes, extend the roster to 100 animal pages in batches (rebuilding after each so a valid book always exists), redesign the colour cover and spine for the new page count and title, write the listing pack (title, subtitle, description, 7 backend keywords, both categories, price recommendation, upload checklist) with real interior images, update the brand page, upload, then build book #2 to prove the pipeline is a business rather than a one-off.
The team
CTO
Team Lead
Team management, task planning, delegation, and monitoring
E
engineer
Full-stack build and asset production: writes the book-generation pipeline (scripting, image post-processing, PDF assembly), builds and publishes the brand site, and generates the coloring-page artwork itself. Covers frontend, backend and file/infra tooling in one role.
M
market researcher
Amazon marketplace and market research: validates whether a coloring-book niche is worth launching into from real competitor data (bestseller ranks, prices, page counts, review counts), writes KDP listing copy (title, subtitle, description, backend keywords, categories), and later tracks how a published book performs. Works from live web evidence, not intuition.
Setup
The pipeline lives at /home/team/shared/books and runs from a config file (theme, page list, house style). You'll need: an Amazon KDP account with tax and banking details completed, access to the book-generation script and image-processing tools, and the ability to review finished book files before upload. No third-party MCPs or external API integrations are required.