priority-based

A routing layer that automatically selects the best LLM per request based on priority flags (speed, cost, quality, balanced) using a weighted score, with under 1ms decision time and built-in fallback, caching, and metrics.