The BabyView Project
The child's view: a frame from a BabyView pilot recording.
The BabyView Project is a research initiative dedicated to capturing children’s everyday experiences via high-resolution videos taken from the infant perspective, and to using the resulting video data to better understand cognitive development. Our dataset is currently the largest open video dataset of children’s everyday experience to date, both in terms of the number of hours and the diversity of the participating families; data collection is currently ongoing.
For researchers
What's in the dataset and how to access it, our data use and ethics guidelines, the camera design, and our publications.
For families
How to take part in BabyView, how we keep your videos secure, and how AI is used in our research.
Data Release Snapshot
So far we have collected 3,898 hours of video from 99 children. See the dataset page for details and access.
| Release | Status | Hours | Children | Released |
|---|---|---|---|---|
| 2026.1 | In embargo | 2,701 | 51 | – |
| 2025.2 | Released | 1,505 | 37 | Jan 2026 |
| 2025.1.preschool | Released | 31 | 26 | Aug 2025 |
| 2025.1 | Released | 965 | 32 | Aug 2025 |
| 2024.1 | Not released | 507 | 24 | – |
Each main release includes all the data in the previous one, so hours are not additive across rows. Hours include all datasets in a release. Updated October 16, 2026.
Funding
The BabyView project acknowledges generous support from:
- The Stanford Human-Centered AI Initiative (HAI) Hoffman-Yee grant program
- Schmidt Futures
- Meta, Inc.
- The Stanford Center for the Study of Language and Information John Crosby Olney Fund
- Amazon, Inc.
- Compute support from the Microsoft Accelerating Foundation Models Research (AFMR) program
- NIH K99HD108386 to BLL