Multi-camera intelligence

Continuous multi-camera detection, no GPU required

Request an Evaluation →
01 // Compression Engine

Specialized Quantization

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.

02 // Stability

Isolated Per-Camera Streams

Every camera runs on its own pipeline, so a troubled stream cannot disturb the others. Under pressure the engine sheds at the door rather than mid-flight, and it has yet to fail a run.

03 // Reliability

Memory-Safe Core

The pipeline is native Rust and memory-safe by construction, with no interpreted layer anywhere in the serving path.

04 // Deployment

Built for Sites Without a GPU

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.

05 // Observability

Production Telemetry

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

12 cameras
720p streams held under detection at once, with no GPU
30 fps
of video sustained on each, with nothing lost inside the pipeline
60 / sec
detections delivered across the twelve
Under 1 GB
engine memory in use, on systems normally specified at 16 GB and up

720p30 streams over RTSP · 29.5–30.0 fps per camera across three 45-second steady-state passes · 48,333 frames delivered, none of them lost between intake and output at any load measured

Measured — the detail tier, where accuracy leads

40.6 mAP
COCO accuracy, identical on both supported runtimes
1.8×
the throughput of full precision, at matched accuracy and matched CPU
13.5 hours
of camera time benchmarked without an engine failure

720p30 streams over RTSP · 45-second steady-state windows · the full-precision comparison measured on its faster runtime