a window into your child's day, not a summary
See what your child actually did, said, and felt at daycare, reconstructed from camera and audio streams you never have time to watch.
The problem
Daycare reports are generic. 'Had a good day, ate some lunch.' You know your child had moments of frustration, joy, or quiet discovery that no one had time to write down. The gap between the lived day and the two-sentence app notification leaves you guessing about their development and emotional state.
What AI makes possible now
Computer vision models now track specific children across hours of footage without storing or exposing raw video. Audio diarisation isolates your child's speech from background noise. A multi-modal agent watches the stream, notes when your child learns a new word, helps another child, or struggles with a transition, then writes a private narrative just for you. The system never shows you other people's children. It never stores video. It reads the room the way an attentive assistant would, then writes the story you actually want.
How it works
- Cameras and microphones already in the room send a live feed to a local edge device that processes everything on site.
- The system identifies your child by face and voice, then tracks only their interactions, discarding all other identifiable data immediately.
- An agent watches for moments of learning, social exchange, or emotional shift and drafts a timeline with short, plain-sentence descriptions.
- At pickup time, you receive a private narrative: what your child explored, who they connected with, and any moments the staff would want you to know about.
The first thirty days
Ship a single-room pilot in one daycare. Run a small edge computer with a camera and microphone already installed. The output is a daily email to three parents containing five to eight timestamped, plain-English observations about their child only. No app. No dashboard. Just the story.
How it earns
Charge the daycare a flat monthly fee per room, not per child. The daycare pays for a tool that raises parent satisfaction and reduces staff time spent writing daily notes. Parents get the narrative free, which makes the daycare stickier and more referrable.
Why now
On-device vision models and speech diarisation run cheaply and privately without cloud uploads. Parents have zero tolerance for raw video access but high demand for meaningful signal. This was impossible when every frame had to be watched by a human.
First customers
Independent daycares in competitive urban neighborhoods where parent anxiety is high and staff turnover makes consistent handover notes unreliable.
The hard part
Trust is earned in inches and lost in a frame. A single privacy misperception, even if technically false, kills the product. You are selling a promise about data handling to parents who are primed to distrust any camera near their child.
Build this with us
The edge processing pipeline and privacy guarantees need careful architecture before a single frame is captured. We build the full stack, run the hardware, and handle every parent and staff concern as part of our managed service. If you run a daycare and want to be the first to offer real transparency without adding work to your team, we should talk.