Moliverse

A living language universe built by educators: children enter cultural stories with AI Mentors, and return because the world remembers what matters to them.

Comment

Inspiration

MoliVerse — Where Great Educators Become AI Mentors

🌱 Inspiration

MoliVerse started with a simple question:

What if a child could meet a great teacher, even if that teacher lived on the other side of the world?

I have always been interested in languages and the way they open a door into another culture. But while working on language education, I noticed something that bothered me: the most meaningful part of learning a language is often the human relationship behind it.

A great teacher does much more than explain grammar. They tell stories. They share their culture. They remember what a student struggled with last week. They know when to push and when to encourage.

But great educators are limited by something very human: time.

A teacher can only teach so many students. A one-on-one language mentor can also be too expensive or inaccessible for many families, especially for children who live far away from strong educational resources.

At the same time, generative AI has made it possible for technology to communicate, remember context, generate content, speak naturally, and adapt to individual learners.

That made me wonder:

What if AI didn't replace the teacher, but helped the teacher reach people they could never reach before?

That became the foundation of MoliVerse.

💡 What is MoliVerse?

MoliVerse is a Human + AI language learning platform where educators can create their own AI Mentors.

Instead of treating AI as a generic chatbot, MoliVerse treats the educator and their teaching philosophy as the heart of the experience.

An educator can bring their:

  • Teaching style
  • Voice and personality
  • Course materials
  • Cultural knowledge
  • Stories
  • Learning activities

into an AI Mentor.

A learner can then interact with that Mentor through conversations, stories, challenges, and real-world scenarios.

For example, instead of simply completing a French vocabulary exercise, a learner might enter a fictional Parisian night market with a French AI Mentor.

They might need to:

  • Order food
  • Ask for directions
  • Meet a character
  • Understand a cultural reference
  • Solve a small problem
  • Have a conversation in French

The language is no longer just something to memorize.

It becomes part of a world.

🤖 Why AI?

We deliberately chose not to build another AI chatbot that simply answers questions.

The key idea behind MoliVerse is AI Mentor.

The AI handles the parts of learning that benefit from continuous availability and personalization:

  • Daily conversation
  • Repetition
  • Personalized practice
  • Immediate feedback
  • Content adaptation
  • Long-term interaction
  • Learning progress tracking

But the educator remains responsible for the parts that require human creativity and judgment:

  • Designing the learning experience
  • Creating meaningful stories
  • Bringing cultural context
  • Developing a teaching philosophy
  • Knowing when a learner needs deeper human guidance

This creates a different relationship between humans and AI.

AI doesn't replace the teacher. It extends the teacher's reach.

🛠️ How We Built It

We approached MoliVerse as both an education product and an interactive world.

1. Educator → AI Mentor

We designed an educator creation flow where a teacher or creator can define their Mentor's:

  • Personality
  • Teaching style
  • Knowledge base
  • Learning objectives
  • Cultural background
  • Conversation behavior

The goal is to make creating an AI Mentor feel closer to creating a character and learning experience, rather than configuring a technical AI agent.

2. Learner → Learning World

We designed learning experiences around scenarios instead of isolated exercises.

A learner can enter a story, interact with an AI Mentor, and learn language through context.

This is especially important for children because imagination and narrative can make abstract language concepts feel concrete.

3. AI → Personalization

We are connecting the Mentor experience with AI conversation and knowledge resources so that the system can respond dynamically instead of following a completely fixed script.

The long-term goal is for the Mentor to remember learning context and adapt future interactions to the learner.

4. Parent → Trust

Because our initial audience includes children, we also designed the platform around parental control.

Parents should be able to understand:

  • What their child is exploring
  • What they are learning
  • How they are progressing
  • Whether voice interaction is enabled
  • Whether interaction with a real educator is allowed

For us, AI education cannot simply be engaging.

It also has to be trustworthy.

🧩 What We Built With

MoliVerse was built through a combination of rapid prototyping, AI-assisted development, product design, and experimentation.

Our workflow included:

  • Figma for product and interaction design
  • Claude Code for rapid development and iteration
  • Gemini for experimentation and AI workflows
  • AI-generated visual assets for characters and learning worlds
  • Knowledge-base / RAG concepts for grounding AI Mentors in educator-created materials
  • Voice and conversational AI technologies for creating more natural Mentor interactions
  • Web technologies to turn the concept into an accessible browser-based experience

One of the biggest lessons from building MoliVerse was that AI made it possible for a very small team to prototype ideas that would previously have required much larger engineering and content teams.

🚧 Challenges

1. The hardest question was: "Why does this need AI?"

Our first instinct was to add AI everywhere.

But we quickly realized that using AI is not the same as having an AI-native product.

A chatbot attached to a language course would not be enough.

We had to ask:

What can AI make fundamentally different?

The answer became the Mentor relationship: continuous interaction, memory, personalization, and the ability for one educator to scale their presence.

2. Balancing AI with the human element

There is an obvious tension in an AI education product.

If we automate too much, we risk turning education into another generic chatbot experience.

If we rely too heavily on humans, we lose the scalability that makes AI valuable.

We therefore designed MoliVerse around a Human + AI division of labor:

Humans create meaning. AI creates scale.

That principle became one of the most important product decisions we made.

3. Designing for children

Building for children created another layer of complexity.

A product can be fun without being educational.

It can also be educational without being fun.

But we wanted to create something where learning, storytelling, culture, and companionship reinforce each other.

At the same time, children's interactions with AI introduce important questions around:

  • Privacy
  • Parental consent
  • Content safety
  • Voice interaction
  • Appropriate AI behavior
  • Human educator involvement

These are not features we can simply add at the end. They have to be part of the product architecture.

4. Building too many things at once

Perhaps our most practical challenge was scope.

We initially imagined a huge ecosystem:

AI teachers, games, stories, avatars, courses, communities, progress reports, parent dashboards, and more.

But a startup does not become meaningful by having the most features.

It becomes meaningful by proving one important behavior.

So we narrowed our MVP around three questions:

Will educators want to create their own AI Mentor?

Will learners want to return to the same Mentor?

Will parents trust this form of learning?

Those questions became much more important than simply adding more features.

📚 What I Learned

1. AI is not the product. The relationship is.

The most important thing I learned is that technological novelty disappears quickly.

People don't ultimately care that something uses an LLM.

They care about what the technology allows them to feel, do, or become.

For MoliVerse, that means the real product isn't "an AI tutor."

It is the relationship between a learner and a Mentor.

2. Constraints make better products

Building an ambitious idea with limited time and resources forced us to prioritize.

Instead of asking:

"What else can we add?"

we started asking:

"What is the smallest experience that proves our idea?"

That changed the way we build.

3. Education is fundamentally human

The more we experimented with AI, the more convinced I became that education should not become less human because of AI.

It should become more human.

Technology can provide access, repetition, personalization, and scale.

But curiosity, encouragement, cultural understanding, inspiration, and trust are still deeply human.

That is why our philosophy became:

AI should not replace the people who inspire us. It should help more people meet them.

🌍 What's Next?

Our current prototype is only the beginning.

We want to test MoliVerse with real educators and learners and understand whether people genuinely develop long-term relationships with AI Mentors.

From there, we envision three directions.

For learners

A personalized learning universe where learners can discover different Mentors, cultures, languages, and stories.

For educators

A creator platform where educators can build, publish, and monetize their own AI Mentors without needing to become AI engineers.

For institutions

A scalable AI Mentor infrastructure that schools and educational organizations can use to extend access to high-quality educators.

Ultimately, we don't want MoliVerse to become another place where people "do AI lessons."

We want it to become a place where people meet the teachers, stories, and cultures that change the way they see the world.

❤️ Our Vision

There are millions of talented educators in the world.

But their impact is often limited by geography, time, and money.

AI gives us an opportunity to change that.

Not by creating a world without teachers.

But by creating a world where a great teacher can reach someone they would otherwise never have met.

One educator. One AI Mentor. Thousands of learners. Endless possibilities.

That is the future we are trying to build with MoliVerse.

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for Moliverse

Built With

  • deepseek
  • heygen
  • tripo
Share this project:

Updates

Submission history