an ai ux platform
Just a wrapper is only an insult when the wrapper is empty. Ours holds what the model never does — your context, your history, the way you work — so changing the model underneath changes nothing you would miss. That layer is the platform, two products already run on it, and the work has a name: AI UX.
the family name
Studinari. Ruminari. The prefix changes with the problem; the ending never does. Nari is Alunori's AI project family, and the ending is a promise about the experience — this one holds your context, carries it forward, and answers you rather than a generic user.
The prefix is the problem being solved. The ending is the promise — context held, carried forward, and answered to the person actually asking.
Every Nari project stands on the same idea: the model is rented, the context is yours. The thinking underneath is shared even when the products are not.
A study workspace and a prompt platform have almost nothing in common on screen. A family is an approach, not a template.
A project earns the ending only once losing it would actually cost you something. That is the bar the name sets, and the whole difference between a wrapper and a replaceable one.
the product
One works on how a person learns. The other works on how a person and a model talk to each other. They share everything you cannot see, which is the whole reason the platform exists.
a study workspace
Plans, practice, and reflection in one system instead of five apps — run by an AI team that already knows the student, so nothing has to be explained twice.
Read about Studinaria better-ai-experience platform
Sharper prompts, and automatic repair of the context a long conversation quietly loses — the two pieces of AI UX everyone is still doing by hand.
Read about Ruminarithe layer we work on
Models get better, cheaper, and more interchangeable every few months, and swapping one for another is a config change. What does not come across with it is everything the tool had learned about you. That is the layer Alunori builds on — and the only reason a wrapper is ever hard to replace.
One of those rows can change on a Tuesday. The other is the reason you would still be here on Wednesday.
A good answer to the wrong situation is still wrong. The room, the history, and the constraints have to arrive with the question — and assembling them is work no model does for itself.
Long conversations drift and forget. Progress should survive the gap between sessions, so nobody rebuilds the setup before getting back to the actual work.
Level, pace, pressure, and goal change what a useful reply looks like. Personalization is not a friendlier tone — it is a different answer.
None of the three is a model problem waiting on a bigger model. They are design problems — AI UX — and every one of them is something you would have to build again from nothing anywhere else. That is what makes a wrapper worth keeping.
the founder
I started building because I needed the thing I was building. Every tool I tried could answer a question — none of them knew what I had been working on the day before.
Studinari came out of my own studying: the planning, the pressure, the slow work of figuring out what to do next. Ruminari came out of what happened while I built it — watching good conversations lose their thread, and re-explaining the same context over and over.
Alunori is the company that grew out of both. People call this layer a wrapper, and they are right — the only question worth asking is whether the wrapper is worth anything on its own. Ours is, because it keeps what the model drops. That is why Nari is a family and not a product.
Joshua Yoon founder, alunori
what comes next
Studying and prompting were the first two places the experience broke down. They are not the last. The same idea points naturally at the rest of a working day — how people keep what they learn, how they decide, how they pick things back up after a week away.
nari
The name is not public yet.
We would rather ship two things you would notice losing than ten you would not. The next name gets announced when it earns the ending.