Live scoreboard · api.statefulai.tech/public/metrics

Real metrics for an operational cognition layer.

Two A/B demos run continuously against the alpha API. Same prompt, same agent, same model — one arm uses ContextOS as operational memory, the other doesn't. Tokens, time, tool calls, and completion rates are captured for every run. Numbers below come from — measured demo comparisons. The dogfood timeline below mirrors this repo's own evolution.

last refresh · 5/28/2026, 9:27:19 AM