Problem
Parents take thousands of photos of their kids, but almost none of the personality, voice, or small moments are preserved. A funny answer, a phase, or how they think at age 6 disappears because recording it requires effort and consistency.
Traditional journaling fails because parents do not have time, and kids will not sit down to write.
Solution
Create a system of short guided prompts and repeatable interviews that turns casual conversations into structured memories.
Tidbit is a mobile app where parents record quick prompted videos with their children. Entries are automatically transcribed, organized, and searchable, and over time compile into a living timeline and yearly memory book.
Key Decisions
- Daily and themed prompts designed specifically for kids
- Repeatable yearly interviews to track growth over time
- Automatic transcription and search so memories stay usable
- AI organization into chapters and recaps instead of raw clips
What I learned
The hardest challenge was behavioral, not technical. Kids acted differently when they could see themselves being recorded.
The product had to move toward invisible capture, like audio-first or parent-controlled recording. I did not yet have the technical depth to fully implement this, but it changed how I design products: start with behavior, not features.
Where this came from
Tidbit started as an earlier concept called InTime, and I kept pushing on the same core problem over multiple iterations.
I first partnered with a Pakistan-based engineering team to build the product. As AI tooling matured, I realized how much more valuable automatic memory structuring could become, and I relaunched the effort by cofounding with two Amazon engineers.