Public GPU capacity across our Livepeer AI endpoints.
FastAgents for Livepeer
FastAgents compute mesh, online and dispatch-ready.
A public, sanitized view of the FastAgents fleet: Livepeer AI compute capacity, online GPU capacity, recent Daydream streams, and BlueClaw/BYOC LLM activity.
Recent Daydream/MCP stream volume from public endpoints.
Unique public-site visitors observed in nginx logs.
Estimated fleet GPU energy use from the 24h power curve.
Site heartbeat
Visitors around the world
Approximate country-level dots from recent public-site access logs. No raw IPs are exposed.
Loading visitor heartbeat...
Daydream video-to-video
Online video AI capacity
Capacity is dynamic and reflects currently advertised lanes.
Output FPS, 24h
--
Fresh-gated runner FPS, averaged across active endpoints.
Hourly jobs, 3h avg
--
Three-hour moving average of rolling one-hour Daydream jobs across active endpoints.
BYOC / LLM
BYOC LLM capacity
FastAgents-hosted model pools with public-safe aggregate usage.
Prompt token volume from completed BYOC requests.
Completion token volume from completed BYOC requests.
Input and output request throughput, scaled to the 24h peak.
Successful and failed routed BYOC requests from endpoint logs.
Loading model list...
GPU inventory
GPU types online
Public-safe hardware summary, grouped by GPU family and VRAM class.
Overall GPU power consumption
--
Current power draw across scraped online GPUs.