01

Children deserve patience.

Every interaction we ship has to feel like a teacher with all the time in the world. If a feature rushes a child, embarrasses them, or hurries them past a moment of confusion, we have built it wrong. Patience is not a personality trait of our products — it is a design specification.

02

Play is cognitive work.

Children do their most serious thinking through play. We design for play, not around it. A product that tries to be educational by stripping out joy has misunderstood what education is.

03

AI should reduce parenting anxiety, not create dependence.

If a feature would make a parent feel less capable on their own, we don't build it. We want parents to put down our products and feel more like the parent they want to be — not more like a user of our software.

04

Childhood is not an optimization problem.

We will not score, rank, or gamify the things that should remain whole. There are parts of growing up that don't get better when they're measured. We try to be honest about which parts those are.

05

Multilingual children are the default, not the edge case.

Most children in the world grow up between languages. We start every product by designing for that child first. Monolingual is a special case, not the baseline.

06

We watch before we measure.

Qualitative observation precedes quantitative dashboards. A child reading in a Bengaluru flat at 9pm tells us more than a usage metric ever will. We don't ship features whose effect we cannot observe with our own eyes.

07

Parents are partners, not users.

We design the things parents see with the same care as the things their children see. A parent dashboard that feels like a SaaS admin panel is a failure, even if it is technically correct. Parents deserve warmth too.

08

We ship when a child reaches for it twice.

Not when the deadline arrives. Not when the roadmap calls for it. When a real child uses our product, puts it down, and asks to use it again — that is the readiness signal. Nothing else qualifies.

09

Data is borrowed, not owned.

We hold children's data the way we'd want our own children's data held. We minimize what we collect. We don't sell. We don't share with advertisers. We delete what we don't need. This is not a compliance posture; it is a moral one.

10

AI must be honest about being AI.

We don't pretend our software is human, and we don't pretend it is magic. We tell children, in age-appropriate ways, that they are speaking to a machine. We tell parents what the machine can and cannot do. The trust we are asking for depends on this.

11

Slow is a feature.

Some lessons take time. Reading takes time. Confidence takes time. A child's relationship with language takes years. We don't compress what shouldn't be compressed. Our products are willing to be slow.

12

We write down what we believe.

These principles exist in writing so that they can be held against us. If we ever ship something that contradicts one of them, someone — a parent, a journalist, a future hire, our future selves — should be able to point at the contradiction. Putting beliefs in public is how we stay honest.

Last updated 25 May 2026. Signed by Anshul, founder.