Thanks to AI & Open Source

This project was built around Ollama. Everything in the app runs on light, reliable models that are free and 100% local — the Ollama models run the agents, carry the whole system, and are what every memory test is run on, right on your machine.

Claude (Fable 5, Opus), Gemini and MiniMax played a big part in building it: every time an agent crashed, a local model became unstable, or I was stuck on a problem, they were there to help me design, debug and fix the code — the best Claude models most of all, especially these last weeks. Grok helped in a different way: a second opinion on the builds I shipped, and advice when I hit something hard. And what ships still runs on Ollama.

Ollama was the heart of the project.

Claude, Grok, MiniMax and Gemini were the firefighters.

With one deliberate exception: everything touching the $LOAM token, the LAND Fund and security information was built mostly with the latest PRO models from Claude. For the critical pieces, you bring in the best tools available.