Skip to content

What is Kaba?

“Kaba” means hippopotamus in Japanese. Or at least sounds like it. Or something. Our logo is a Hippo. Hippos are cute. They are also extremely territorial. They are among the most dangerous animals with respect to deathly human encounters each year.

So, cute but territorial seems about right.

Approachable, because you don’t need to be a PhD AI researcher to use Kaba. Electric cars were a different experience at first in comparison to internal combustion engine automobiles (very quiet, no gears shifting, a large center console screen as opposed to knobs & dials), but they were still a car. Just get in and drive. Kaba is the same way. It has its quirks and features, but fundamentally, it’s kind of like a browser (but of course, it’s much much more). Just get in and drive.

Territorial, because much like the Hungry Hungry Hippos boardgame known to kids that grew up in the 80s & 90s, Kaba is deathly serious about data, privacy, and security. We help you to gobble it all up, make sense of it, and, eventually, train on it. This is in service of anyone being able to have and build their own AI. Individuals, families & groups, business, enterprises, or even entire municipalities and countries.

Here is a Kaba Enclave, a personal box running Kaba. Very cute, and yet, it houses our code, our data, our corp chat, so it’s territorial.

Abstractly speaking, Kaba is your:

Compendium of context

Library of language

Axiom of action

Depot of data

Arsenal of attention

Scriptorium of synthesis

Kaba is everything you need to scale your observable universe and the tooling and pipelines that prep that data and context for training. This is all in service of allowing anyone to build their own library of models.

Concretely speaking, it is a new kind of operating environment where you can learn, work, and play as you normally would. It launches just like an application, but houses an entire world — runtime(s), browser, terminal, file explorer, and pretty much anything else you need to do what you normally do.