The process of trying and testing each model for each purpose has been a little painful, particularly when working in the context of a microapp.

To simplify this, I started working on a library so I can easily switch prompts between LLMs and assess the return values.

I've now published this as an open source project called Quolo.

Keen to keep building this publicly, but would like some feedback before I go too far down the rabbit hole.

I appreciate my experience is not everyone's. Would love your feedback on the idea, and any ideas you have about what it could turn into over time.

I'm really excited to keep evolving this.

Thank you.