Please let me know what you think :D

Super curious to hear feedback or suggestions. We have a lot of different directions we're going to explore, the next big things we're looking at are:

Model Router - Routing between different models seamlessly without killing cache rate

Incremental Progress - Add the ability for the AI system to tweak tools/prompts seamlessly to improve cost/efficiency/response

Skill / API Aggregation - Avoiding the MiTM route and using a skill so the AI reports its usage. The idea is that this would enable it to work with things like Claude Review, etc...