[tstone86]

This would be great, and feels like the direction things will quickly head next. How can I break down larger jobs to tasks, with specific portions going to the public LLMs, and the more sensitive private data tasks going to perhaps a self hosted model, all while being able to monitor accuracy. Love it!