I built a voice-first family timeline app and need blunt feedback
I have been building Hearth, a private family timeline app for my own family, and I would love blunt feedback from other builders.
The problem I started with was pretty ordinary: family life gets scattered everywhere.
For us, the pieces were split across Google Calendar, shared photo albums, text messages, notes, packing checklists, and random grandparent updates. Memories and logistics were technically saved somewhere, but never in one place that told the actual story of family life.
The product idea is a shared family timeline where past memories and upcoming events live together.
The core feature is voice capture.
Instead of opening a form and manually organizing everything, you can say something like:
"Luca had his 18-month checkup today. He is 24 pounds, doctor said speech is on track, next appointment is in August."
Hearth turns that into a structured timeline entry: health note, date, summary, and maybe a future reminder.
Or:
"We are going to Yosemite next weekend. Need to pack jackets, snacks, the carrier, and book the dog sitter."
That becomes a trip entry with a checklist and countdown.
We have been using the v1 in our family, and it has genuinely changed how often we capture small moments. The surprising thing is that voice makes it feel less like maintaining a database and more like just telling the family record what happened.
Where I am stuck is positioning and trust.
This is sensitive data: kid milestones, health notes, photos, voice transcription, maybe calendar context. In v1, this probably cannot be full end-to-end encryption because voice transcription and AI extraction need to process the entry content. The honest model is: private, encrypted in transit and at rest, family-scoped access, signed media URLs, not used to train models, but not "the server can never see anything."
I am trying to figure out:
Is this a meaningful enough problem, or just a nice-to-have?
Is voice-to-entry the real wedge here?
Would parents trust this if the privacy model is explained clearly?
Should there be a non-AI / E2EE mode from day one?
As a side project, what would you test next: parent interviews, a tiny beta, a landing page, or more usage inside our own family?
No link or signup ask here. I am mainly trying to avoid fooling myself because my own family loves it.