VocalPilot does what an engineer does with a hand on the fader: it listens to the vocal and evens out phrase loudness in real time. Quiet lines come up, shouted lines come down, and it all happens smoothly — no pumping, no jumps and no noise being dragged up. The compressor after it no longer has to squash a 10 dB gap between a whispered and a shouted verse, so the vocal sounds natural and still sits steady in the mix. Built for rap, trap, drill and melodic vocals — verses, hooks, ad-libs and backing vocals. Auto Target is on by default: the target is the average level of the voice over the last ~8 seconds, so the overall loudness of the track doesn't change and only the phrases are evened out against each other. On an uneven take the loudness difference between phrases drops from about 12 to about 4 LU. Want your own level? LEARN or dragging the target line on the display switches to a manual Target. The detector hears the voice, not just level: breaths, sibilants, clicks and background noise are ignored (it recognises vocal voicing), short events under Min Event never move the fader, and in pauses the fader parks at the average gain — the breath after a quiet line doesn't get that line's boost. Phrase Mode follows phrases, not syllables: the gain settles at the start of a line and stays there instead of chasing every word. The loud-word guard instantly catches a shout after a quiet phrase, and a boost safety never pushes a hot take into clipping. Vocal Type (Verse, Hook, Ad-Lib, Backing) tunes the windows, phrase hold times and loud-word headroom to the role of the track. Auto Target, Target, Max Boost (9 dB by default), Max Cut, Amount, Speed, Smoothing, Sensitivity, Min Event and Lookahead (0–150 ms; 0 = zero latency for tracking). The display shows the history: input, output and the rider's work area (lifted / lowered), a draggable target line, phrase bars with each phrase's gain, ignored breaths and the gain curve. The RIDE module is a motorised fader with a draggable range and IN / OUT meters in LUFS. Bypass with Level Match compares the dynamics alone at equal loudness.