prompt amized is a toolkit to help people write to evaluate and score and understand while they iterate on their while doing prompt engineering. We have a big corpus that luckily was contributed by I forgot which university but a bunch of PhD people did a text-to-sequel contest at Yale. So for us we're able to take that test set and some of our own test sets and run it at scale against you know open AI or against llama or against different models with 73 percent accuracy. It's non binary so it's also you don't need perfection to ship you just need better than the previous version or good enough to start it with.