







Teller Family
Turn family photos into illustrated bedtime stories with AI.
Hypothesis
Personalized content is the next evolution in media. Media has gone from edited and linear, to recorded and saved, to recommended and shared. Personalization algorithms will give way to personalized content — content that features the viewer and their world. I'm not sure how much space it will take up, and professionally-created content will always dominate, just as linear still has its place today, but personalized content will continue to grow.
Timeline
3 weeks
The Problem
Parents want fresh stories to read to their kids. Kids love seeing pictures and illustrations of the world they know and live in — their pets, their house, their family. Most AI story apps generate bland, generic stories.
The Solution
Teller Family lets parents upload family photos and generates fully illustrated, personalized bedtime stories. Each story features the child's real world reimagined through AI-generated illustrations matched to the narrative. In addition to uploading photos, parents can choose a reading level and a genre for the story. The genre effects the style of the illustrations.
How AI Is Used
A two-stage AI pipeline powers the experience. First, parents upload family photos that are sent to fal.ai to generate stylized illustrations — transforming real photos into story-ready artwork. Those illustrations are then analyzed by Claude to craft an age-appropriate narrative that weaves the illustrated scenes into a cohesive bedtime story. Once illustrations are generated, the original photos are deleted — they're never stored, in an effort to protect family privacy.
Tech Stack
Built as a cross-platform mobile app with a serverless backend and AI services orchestrating the story pipeline.
What I Learned
Getting people to adopt a new app or tool is hard! My problem statement and hypothesis were based on real conversations and anecdotes. But that doesn't guarantee usage! Seeing this firsthand from the ground up has been enlightening.
Aside from the perils of consumer app launches, I learned how AI sees and processes images, and how it can take the safest path to accomplish its goal. Both of these can sometimes cause the AI to not re-generate the image as an illustration. This project gave me a much deeper understanding of how AI works.