Continuous multi-camera detection, no GPU required
Request an Evaluation →Detection models are compressed by our own quantization system, which holds the accuracy that general-purpose tooling gives away. That margin is what lets a full detection workload run with no GPU at all.
Every camera runs on its own pipeline, so a troubled stream cannot disturb the others. Across our entire benchmark programme the engine has yet to fail a run.
The pipeline is native Rust and memory-safe by construction, with no interpreted layer anywhere in the serving path.
Cryphex is tuned to run efficiently where there is no accelerator to fall back on. Where a site already has one, it will use CUDA or DirectML natively.
Prometheus and REST monitoring are built in, with per-stage latency and overload alerting, so an operator sees a developing problem before a customer reports one.
Measured — the density tier, on one mid-range device
720p30 streams over RTSP · 45-second steady-state windows · repeated across three interleaved passes
Measured — the detail tier, where accuracy leads
720p30 streams over RTSP · 45-second steady-state windows · the full-precision comparison measured on its faster runtime