Trust
Two things sit under every cue — one you drive, one that runs on its own. Both are built from the same infrastructure the rest of the portfolio runs on. Neither one judges you.
Every cue has three settings, and you slide between them per reaction. The cue file never changes — only who can see it does. Private is the default for every scrobble and every unfinished thought.
Two mechanical layers, no human review queue, no interpretation. It measures the audio — it never decides whether you're angry.
The waveform is already in the file. We check it against a declared physical boundary — a sustained level held too long routes to review. It's a measurement, not a judgment. Nobody in voice is running this.
How it actually works
Both floors reference a declared agreement — the platform's boundaries, stated up front, so you know the rules before you ever hit record. No black box deciding your fate after the fact.
Whisper produces the text; the text is checked against a declared agreement before a cue goes public. Mechanical, not a review queue. what was said → the rules → clear or hold
The amplitude envelope is already extracted for the player's waveform. A declared threshold reads it: sustained level past a boundary routes to review. the waveform you already have → a number → clear or hold
Control that measures instead of judges is testing now.
Join the tester list