Aramika

Aramika is a live gamified platform for mastering Talmudic Aramaic. AI agents power teacher reports, production operations, and B2B lead qualification for a real revenue-generating business.

Comment

Inspiration

Aramika started with a different app I built for myself to stay consistent with my learning, using gamification, streaks, and rewards.

When I showed it to a friend, I joked, “Look, I built Duolingo for Talmud study.” He said, “Nice, but that’s really not Duolingo.”

And then it clicked: what if I actually built something like Duolingo, but for Aramaic?

I have studied Talmud for years and had almost forgotten how difficult it was at the beginning to keep running into words you simply do not understand. Suddenly I saw that language barrier again as a problem I could try to solve for learners across a global Jewish community of millions.

What Aramika does

Aramika teaches Talmudic Aramaic through a 66-level path with about 300 words, exercises, stories, leagues, streaks, and rewards. It is available in five languages and includes a school dashboard where teachers can track student progress and difficulties.

During the hackathon, I added three AI workflows that now run in production:

Teacher Reports analyze real learning data and return class insights, while student identities are pseudonymized before data is sent to Gemini.

Operations Agent analyzes production data, compares periods, and returns a structured decision — monitor, create an operations task, or escalate a critical issue. The backend validates and executes the allowed action.

B2B Deal Desk uses Gemini and Google Search to research institutions, find evidence, score fit, and route leads into Priority, Review, or Reject. For relevant leads, it prepares a personalized email; I approve it before sending.

How I built it

I am a 20-year-old solo founder. I cannot write a single line of code on my own.

I built Aramika from start to finish using Google AI Studio, with Gemini as my teacher and partner.

When I started, terms like Firebase, databases, servers, Stripe, and RevenueCat were new to me. As the product grew, I realized that even if AI writes the code, I still need to understand how the system works, how the parts connect, how to investigate bugs, and when an AI-generated solution is simply not good enough.

Today there is a real app in production with users, payments, subscriptions, and AI systems — and I still have not written a single line of its code myself.

Challenges

One of the first challenges was design. I am very sensitive to aesthetics, and the early versions looked generic and almost screamed “made by AI.” At one point, I put the project aside.

Later I read about a father and son in London who had built an AI-powered football video app and reached thousands of users. It probably hit my ego a little — I thought, if they can do it, there is no reason I cannot. I came back to Aramika determined to make it work.

There was also a technical bug that took almost two days and drove me crazy: a customer paid through Stripe, but the subscription reached RevenueCat as anonymous, so the system did not know which user should receive Premium. After hours of testing, I found that the user’s UID was not attached to the metadata correctly — one tiny detail that broke the whole chain.

I also had to learn when to ship. As a perfectionist, I could always see one more thing to improve. I had to accept that the product could be far from my final vision and already be useful.

What I’m proud of

I am proud that I simply kept going. This is the biggest project I have ever taken on, and I built it alone over several months in areas I had never worked in before.

I only started meaningfully distributing Aramika organically in the past week. In seven days, Google Analytics recorded 587 new users and 290 active users across 27 countries, with an average engagement time of 21 minutes and 10 seconds per active user.

During those same seven days, RevenueCat recorded $550 in revenue from 27 transactions. Aramika currently has 34 active subscriptions and about $94 MRR.

All of that came with $0 spent on paid advertising.

But what excites me most is not the numbers. I heard about a student who had struggled with Gemara and went on to learn hundreds of Aramaic words. And in my younger brother’s class, where the teacher ran an Aramika competition, students started quizzing each other on Aramaic words even during breaks.

That means more to me than any graph.

What I learned

I learned that I can take on something much bigger than anything I had done before, stick with it for months, and keep moving even when I have no idea how I am going to solve the next problem.

I also learned how much AI can open up worlds that used to be closed because of technical barriers.

A solo founder can now take an idea from zero to a real product without a development team or raising money just to start building. That changes the starting point for entrepreneurs.

I cannot write a single line of code — but I built a real product in production.

In a way, I was a solo founder with one developer: Google AI Studio.

What’s next

Aramika is coming to the Apple App Store soon. It is already available on Google Play, and my Apple developer account is ready.

Shortly after Aramika started gaining users, Chinuch App — a technology platform for Jewish schools — reached out to me directly and proposed integrating Aramika through its SDK, including the possibility of offering Aramika’s institutional tier through their platform. Those discussions are already moving forward.

I want to turn Aramika into a complete system for learning Aramaic: hundreds of levels, conjugations, sentence structure, real passages from the Talmud, far more content, and design and animation at the level of the best learning apps in the world.

My goal is to put the first revenue straight back into the product — into professional content, design, and building the version of Aramika that I have in my head.

Built With

Share this project:

Updates