Problem
Arabic-speaking parents had few good options for calming, age-appropriate bedtime content in Arabic, and producing that volume of stories, narration, and illustrations by hand doesn't scale.
Role
Full-Stack Developer. Built the platform end-to-end, including the AI pipeline that generates every story's text, audio, illustrations/video, and personalized book pages.
Stack
Decisions
Built with Next.js for RTL-first Arabic support along with SSR/SEO so story pages are discoverable.
Built an AI generation pipeline that produces the story text, narration audio, and illustrations/video for every tale, instead of commissioning content manually.
Organized the entire catalog by age band (0-2, 2-4, 4-6, 6-8, 8-10, 10+) so parents can quickly find age-appropriate stories.
Used MinIO (S3-compatible object storage) on Railway to serve the generated audio and illustration/video assets behind signed URLs.
Extended the same generation pipeline to produce personalized printed books, swapping in the child's name and face throughout.
Mixed free and subscription-gated stories so the app has a real upgrade path instead of an all-or-nothing paywall.
Outcomes
Launched at hadoota.net with a full library of AI-generated, narrated, age-categorized Arabic bedtime stories.
Shipped a personalized printed book flow where the AI pipeline generates a physical storybook with the child's name and face in it.