.Wave

Summary

.Wave provides dedicated inference capacity for real-time workloads, alongside serverless streaming models for speech and conversation. Our Dedicated Lanes give you the guarantees of a dedicated endpoint (latency, capacity, locked-in price) without paying for idle GPUs. Capacity is matched to your traffic on an agreed schedule: more during peak hours, less during quieter periods, or none when you don’t need service. You only pay for the capacity and time windows you reserve. Our .wave Live API, powered by our WPK inference engine, provides streaming speech recognition and full-duplex voice.