Transcode · ingest & encode

Any feed in. Every format out.

Take UDP, HLS, HTTP, SRT or RTMP in and put HLS, DASH, MPEG-TS, UDP multicast, RTMP or SRT out — transcoded on your own GPUs, not a per-node broadcast-encoder licence.

GPU pipeline

• NVDEC / NVENC — hardware decode + encode (H.264, H.265, MPEG-2 in; H.264, H.265 out).
• CPU fallback — libx264, libx265, libsvtav1 when no GPU is present.
• Zero-copy — direct GPU-memory transfer when the same GPU decodes and encodes.
• Cross- & multi-GPU — PCIe transfer via hwupload_cuda; up to six GPUs with NUMA-affine CPU core allocation.

Every stream, one dashboard

Each stream carries multiple inputs and multiple outputs, with live state (running, starting, stopped, error) — the whole line-up under load, on one screen.