Why this exists
Serious research is a loop. You collect material, reason over it, run models on it, and keep what you learned so the next question starts further along.
Every time I tried to do that with AI, I ended up gluing together tools that knew nothing about each other. A chat that forgot last week. Data stuck in cameras and spreadsheets. Models I couldn’t run where the data lived. Answers nobody had argued with.
I trained as a designer, not an engineer. I studied multimedia design, and I taught myself the rest by building, one piece at a time.
So I started building the pieces, one at a time. Pentool, because agents needed a way to draw a figure. Taraference, because some jobs should never leave your own GPU. Kest, because one agent agrees with itself and a team does not. Argus, because the physical world doesn’t arrive as a PDF. LSI, because findings should outlive the conversation that produced them.
TaraSpace is where they meet. The Platform is the newest piece: one key for the models behind all of it.
I’m one person, so I can’t win by sounding bigger than I am. What I can do is show the work. The source is public where I can make it public, every release is dated, and where something isn’t ready, this page says so.




