Smart Routing
Starchild routes each request to an appropriate model automatically. Good results without paying for more than the task needs.
How it works
Each request is matched to a model based on complexity and type. A simple question gets a fast, cheap model. A nuanced analysis of a long document gets a stronger one.
What it saves
Without routing, most people default to the most capable model for everything and overpay 5-10x on simple tasks. Routing eliminates that.
Override when you want to
Switch models manually in session when you want a different output style or know a specific model performs better for your task.