Turn a short story into an illustrated, narrated storybook with consistent characters using five Claude Code skills
A free tutorial and open-source skill set that splits a story into scenes, builds a character reference sheet, generates each scene with the earlier images as references, adds narration and publishes one HTML file, for about two dollars per ten scenes.
Evidence: The source shows its work: steps, screenshots, code or data you can inspect.
The business problem
Creators making illustrated stories or children's books with AI image models find that characters change appearance from page to page, and stitching images and narration together by hand is slow.
What was tried
1. Write a short story. 2. In Claude Code, run an orchestrator skill that calls four others in order. 3. The scene-splitter breaks the story into scenes. 4. The illustrator writes a visual bible describing every character and location, generates one reference image per character, then generates each scene through the Fal API, passing in the character references and the previous scene's image. 5. The narrator creates audio for each scene with ElevenLabs v3 through its MCP server, using emotion tags. 6. The publisher pairs images with audio and outputs a single self-contained HTML storybook. For long stories, regenerate reference images at the start of each chapter to limit drift. Image model options named: Nano Banana 2, Nano Banana Pro (Gemini 3 Pro Image) and Seedream 5.0 Lite.
What was reported (not reported)
The author shows one ten-scene demo story with three recurring characters and side-by-side examples with and without the method. He estimates roughly 1.50 to 2.50 dollars per ten-scene story (the FAQ says 1 to 3 dollars). No measured consistency scores, sales or reader results are reported.
Limitations
Vendor: the author sells AI development courses on the same site. One demo, no quantitative test of consistency, and the cost ranges differ slightly within the page. Characters can still drift over long stories. Needs paid Fal and ElevenLabs API keys. Commercial use depends on each model's licence terms, which the author says to check. Google's free Gemini Storybook is faster for one-off stories.
What you need
Claude Code, Fal API key, ElevenLabs API key, and the author's free claude-image-generation repository.
Check the image and voice models' commercial licence terms before selling the output.
Sources
- LearnWithHasan (Hasan Aboul Hasan) ↗ Vendor case study, published July 8, 2026
Source published: July 8, 2026. Last reviewed here: November 1, 2026. Spot a mistake? Tell us.
Tools in this workflow
- Claude Code: Runs the five storybook skillsTry Claude Code
- ElevenLabs: Narration through its MCP serverTry ElevenLabs
- fal: API for the image modelsTry fal
Get new workflows like this by email
A short selection of the most useful new workflows, including the ones that did not work. Free, and you can unsubscribe at any time.
Related workflows
Before fixing a zero-view marketplace shop, measure the base rate: an AI agent pulled 410 competing listings through the public API
An AI agent running a tiny template business on a 200 euro budget had zero sales and almost no views. Instead of rewriting listings again, it sampled 410 competing listings and found that new listings in its niche normally get about three views.
Set up Claude Code for a real SaaS codebase: browser access, allow and deny lists, tests after features and a user-focused system prompt
A bootstrapped founder who has built his platform almost entirely with Claude Code for six months shares the setup he relies on: Chrome access, permission lists, a completion loop, isolated tests and a system prompt that describes the ideal user.
Post-mortem: AI coding agents built 14 apps in seven weeks, then the landing page got zero clicks on its main button
A solo developer used AI coding agents to ship a 14-app suite with over 1,000 commits, then found that two teams used it for free, 78% of visitors left after one page and nobody clicked the main call to action. He narrowed to one product.
Make a launch film for your product with Claude Opus 5.5 writing frame-by-frame HTML, rendered by Playwright and ffmpeg
A student made a 60-second product film for a small web tool by having Claude Code write one HTML file where every frame depends only on time, screenshotting each frame with Playwright, and reviewing contact sheets before the full render.