Real writeups from our engineering team — what we're building, what broke, and what we learned. For students, teachers, and engineers alike.
The Zeretis solver runs in 30 microseconds on a standard laptop — 99.9999% less energy than GPU inference. This post explains the architecture behind that number and why it matters for users, the environment, and the future of math tooling.
Read article →Most tools require you to learn their syntax before you can use them. Here's how we approached removing that barrier — and why it matters for students and teachers.
The difference between symbolic computation and AI approximation — and why it matters for getting exact answers you can trust.
How to set up student accounts, assign problem sets, and use the step-by-step solutions as a teaching tool rather than a shortcut.
A look at the architectural decisions behind the Zeretis backend and why correctness guarantees were non-negotiable from day one.
A complete reference for entering fractions, radicals, integrals, matrices, and more using standard LaTeX notation in the Zeretis solver.
Matrix operations, improved equation parsing, faster step rendering, and a new keyboard shortcut reference panel. Here's everything that shipped.