◈
Generation performance
Separate queue time, generation, upload, and commit. Compare measured latency and throughput by model, with image and video views.
One place to follow your GPU fleets, understand image and video generation, and inspect the evidence behind every run.
Company sign-in required · coredigits.com accounts only
Illustration · no live fleet data shown
Separate queue time, generation, upload, and commit. Compare measured latency and throughput by model, with image and video views.
See GPU utilization, VRAM, temperature, and power alongside agent freshness and worker signals. Spot silence as well as failure.
Inspect traces and recorded evaluations. Track collection gaps and verified archive receipts, even after an ephemeral fleet is gone.
Missing measurements stay missing. Coverage is visible. Iris keeps execution timing separate from generation timing, and spend stays unknown until billing is connected.
Sign in to inspect your fleet and the work moving through it.