[calyhre]

Super interesting. We've been experimenting with promptfoo[1] at my work, and this looks very similar.

[1]: https://github.com/promptfoo/promptfoo

[artas728]

Thanks! While promptfoo offers great strict metrics, Spelltest differentiates by allowing custom prompt-based metrics. This lets you measure unique attributes like empathy or creativity in dialogues and completions.