Your family. Your data. And exactly where AI fits in.

Yes — but in a few clearly defined places, not as some invisible layer running through everything. In Nest, AI helps us write, translate, sort and create Teddy characters. It does not make decisions about your child.

• Daily story: the teacher writes the actual notes. Anthropic Claude turns your child's first name, age and those notes into a readable story — never using photos. Teachers normally send the story themselves. If a draft is still waiting at 18:30, Nest sends it automatically. This runs for every group and does not have a per-child switch.
• Translation: if you enable it, Anthropic Claude automatically translates messages in that chat, including messages from teachers and other parents. It is off by default.
• Urgency check: when you send a chat message, Anthropic Claude reads it once to decide whether it looks urgent, so a teacher can get an immediate alert. That check never writes back to you. The only AI that ever replies in Nest is Vera, our assistant — and when it is Vera, her name is on the message, so you always know who you are talking to.
• Teddy Guard & Teddy Kid: only when you choose these features does OpenAI process photos of your child to create their Teddy version.

A few internal staff tools also use AI for summaries, such as turning a registration form into a shorter overview. Want the complete list? It lives on the AI page in Nest.

Not when the photo is taken. Not when it is uploaded. A normal photo goes straight into secure storage.

AI becomes involved only when you switch on Teddy Guard or create a Teddy Kid character. In that case, OpenAI receives photos of your child — including existing photos in Nest and new photos from that point onward — to create their Teddy version.

The first sticker that covers a face is created on our own servers without AI. Finding your child in new photos happens locally on your own phone. Your facial data stays on your device; only the photo reference and the location of the face are passed on. The recognition model itself is downloaded from public model libraries such as Google and jsDelivr. Your photos are not uploaded to those libraries.

No. Three words often get mixed together here, so let's separate them:

• Processing — the provider receives information to perform the task you asked for.
• Retention — how long that request may remain with the provider. Requests can be kept briefly for abuse and safety checks, for up to 30 days, and are then deleted.
• Training — using our data to improve the provider's AI models. Anthropic and OpenAI do not train their models on our data.

The heart of Nest — your child's data and photos — is stored with Supabase in Zurich, Switzerland.

Around that core are a few specialist services: Vercel delivers the app itself, Resend sends login emails, and Apple and Google deliver push notifications to your device.

And one thing you won't find hiding in Nest: advertising or third-party analytics trackers.

Access follows the job, not curiosity.

Teachers can see the children in their own groups and location. Managers can see their location. The owner (Artem) and our Head of Teddy Operations (Antonella) have full access.

Sensitive numbers get an extra curtain: staff only see the last three characters of bank account numbers, while only Artem and Antonella can view the complete number. BSNs are masked on staff screens as well, with only the final digits visible.

Your personal login link can be used once and expires after 72 hours. Prefer a code? The 6-digit email code lasts 10 minutes.

Because access starts through your email account, keeping that account well protected is an important part of protecting Nest too.

And we're adding another lock for parents who want one: optional authentication through an authenticator app.

No panic button required.

Log in on another device and choose Log out everywhere & get a new link. Every other Nest session stops immediately.

Each phone or computer gets its own session. A session expires after 3 months without use and, even if you use Nest every day, after a maximum of 12 months. Then you simply log in again. A guardian can have up to 10 active devices.

No. Nest works out who you are from your login and checks whether you are actually linked as a guardian before showing child information. If that relationship isn't there, the server refuses the request.

Some spaces are intentionally shared — for example location chats or parent committee chats — so the other families participating in those chats can naturally see what is posted there.

Yes.

Normal photo taking and uploading does not involve AI. The daily story never receives photos either.

Photo AI is used only when Teddy Guard is switched on or a Teddy Kid character is created. Leave those features off for your child, and ordinary photos stay ordinary photos.

This is probably our favourite question.

A change cannot simply be pushed into Nest. It first goes through a pull request and must pass type checks, automated tests, secret scanning and AI code review.

Those reviews come from three independent AI providers — Claude (Anthropic), Codex (OpenAI) and Grok (xAI) — each checking the change on its own. Artem, the owner, orchestrates all of it and carries full responsibility. There is no separate human review team hiding behind that process.

And Nest is not a first attempt: before Teddy Kids, Artem spent seven years building apps, including running a 15-developer office in India from 2017 to 2021 — Zaycare / eKidz, for those who remember. AI lets that experience move a great deal faster.

If a check finds a problem, that release stops until the problem is fixed.

And one layer we still want to add: we have not yet completed an independent external security test. It is on our list.

For anything about privacy or your data, contact Antonella Falcone, our Head of Teddy Operations, via info@teddykids.nl. You can also ask your location manager — real humans are very much still part of Nest. 😉