You.com Inference Router offers multiple models through one OpenAI-compatible API and fails over to partner providers when its own inference capacity is unavailable or full. The company describes zero data retention for model calls and web searches, US hosting and a minimum 99.9% uptime SLA. It also provides model-specific token prices and per-key spending caps. The SLA is a contractual commitment backed by service credits, not measured uptime or a guarantee of uninterrupted service; response quality and latency should be tested on real workloads.