HeyOtto!

As parents, we watched our children grow curious about generative AI—wanting to ask endless questions, create stories, generate images, and explore ideas in ways that felt magical to them. We were excited by their enthusiasm, but deeply concerned about letting them use existing tools. Without proper context, these models often respond in ways no responsible adult ever would: engaging in topics, tones, or depths of discussion that would be considered unthinkable in a real-world conversation between an adult and a child.

That gap became the core problem we wanted to solve. Most AI chatbots treat all users similarly, lacking the awareness and boundaries that any caring adult naturally applies when speaking with a child. They have no understanding of the family’s values, the child’s age and maturity, or which conversations are best left to parents. As a result, these interactions often drift into areas that are confusing, overly mature, or simply not aligned with how parents want their children guided.

We created Otto to change that. Our guiding principle is simple: an AI interacting with children should converse with the same care, respect, and appropriateness we expect from a responsible adult. To make this possible, parents begin by creating a family profile that defines their core values, cultural background, religious beliefs, and specific content boundaries. This framework, combined with each child’s profile—including name, age, and relevant background—gives Otto the context it needs to respond thoughtfully and stay within appropriate limits.

The safety system reinforces this approach: built-in moderation and alerting addresses general risks and ensures parents are notified when any conversation or interaction becomes concerning. Through a dedicated dashboard, parents maintain full visibility and can refine boundaries, adjust profile settings, and further tailor their child's experience at any time.

Our process evolved as we built and tested with early families. We started focused on strong filtering and compliance (designed COPPA-safe from the ground up), but feedback showed that true appropriateness requires more than filtering—it requires positive guidance shaped by context. We deepened the personalization system to weave values directly into responses, expanded creative features like collaborative storytelling and AI art to keep children engaged, and refined the alerting system to feel proactive rather than reactive. What began as an effort to make AI safer evolved into building an AI that genuinely understands and respects the adult-like responsibility owed to children.

The result is HeyOtto!—created by parents who wanted an AI companion they could trust completely with their own kids. We are excited to share it with the Product Hunt community and welcome your thoughts as we continue to improve it.