Decision Models · demos
Demo 1 · speed · s1-fast
Every chat message moderated before it hits the screen.
A live stream chat gets dozens of messages a second. Here, each one gets a real moderation decision from s1-fast on api.decisionmodels.io (allow, hide spam, hide abuse or escalate to a human) before it appears. Press the button and 60 messages arrive at once.
Decisions / second
–
Latency p50 · round trip
–ms
Latency p95 · round trip
–ms
All 60 decided after
–ms
s1-fast · live0 / 60
Jev 1.14.0 · recorded replay0 / 60
Same kind of burst (60 messages, 12 in parallel, same question) sent to Jev 1.14.0 from the same server on . Replayed here at its measured timing. No live Jev calls are made from this page.
Your turn · one decision per keystroke
Type a chat message. It is re-moderated on every keystroke.
waiting for you
Each keystroke sends the whole message to s1-fast. If you type faster than decisions return, only the newest text is sent.
Measured, not claimed