You can continue working on any model, seamlessly
As the token subsidy era comes to an expensive end, I started to worry about having all my projects tied to a single model that keeps getting more expensive. Or, finding out that I am stuck with one, while the others get more powerful. Different models are good at different things. but they don't communicate.
I basically wanted to be able to jump into any model at will without losing any work. But I didn't want to have all my information sitting in someone's cloud. So I built basically a "context os" that unifies your context/memory from various tools, updates as you work, and never leaves your control (or your side).
It's basically a bunch of plain text files with instructions for the LLM about how to use them. That's really it.
Free. No sign up. Private. Update itself. You own the files.
And it's free... and totally private.