Learns the room’s fingerprint
Builds a live picture of how energy leaves the channel and finds its way back.
AU for Apple silicon · In development
More vocal.
Less room.
No howl.
FedUp is a zero-algorithmic-latency feedback controller for live sound—built to get meaningfully more level from the mic without hollowing out the voice.
+12 dBgain-before-feedback target
0 samplesalgorithmic latency target
16+ channelsApple-silicon baseline target
MIC INPUT
LOOP CONTROL ACTIVE
CLEAN OUTPUT
THE MIDDLE PATH
Smarter than a feedback destroyer.
Lighter than full AI separation.
Classic boxes wait for a frequency to run away, then leave a cut behind. Heavy AI can clean the room, but at a serious compute cost. FedUp is built for the space between them.
INCLUDED–$149Static notchesFast, reactive, tonal scars
$49 · FEDUPLoop-aware DSPAdaptive, scalable, live
$499AI separationDenoise + dereverb, compute hungry
FEDUP CORELoopSense™PROPRIETARY CONTROL ENGINE
THE PART WE WON’T DIAGRAM
It doesn’t chase loud.
It recognizes the loop.
Feedback is not just a frequency. It is a relationship between a microphone, a speaker, a room, and time. LoopSense watches that relationship—not merely the level meter.
Separates a note from a runaway
Looks for the behavior of a loop before it commits to audible correction.
Uses only the force required
Protection rises with risk, then gets out of the way when the system is stable.
The exact signal model, confidence logic, and control strategy stay in the box. What comes out is simple: more usable vocal, less acoustic fight.
THREE LAYERS. ONE JOB.
01
Cancel the loop
A causal adaptive model learns the acoustic return path and reduces loop energy before it becomes a ring.
02
Control the resonance
Dynamic bands follow unstable frequencies, then get out of the way when the room settles down.
03
Catch the scream
A surgical emergency notch bank remains armed for anything the first two layers cannot contain.
THE BUILD STANDARD
Designed for the show,
not the spec sheet.
LATENCYZero algorithmic delayTypical 5–7 ms real system round trip
FORMATAudio UnitNative Apple silicon, mono + linked stereo
BEHAVIORFail safeInstant bypass, no surprise gain, no permanent tone shift
PROOFMeasured in a real loopGBF, peak CPU, missed deadlines, and vocal damage
“The best feedback processor is the one you stop hearing.”
No denoise theater. No mystery latency. No claim we cannot reproduce. FedUp is being built in public around one hard measurement: how much more usable vocal level can we get before the room takes over?
OWN THE MIDDLE
Pro feedback control.
Working-musician money.
FedUp is not a stripped-down AI cleaner and it is not another automatic EQ. It is focused, loop-aware protection for the people who need the mic louder tonight.
- Perpetual license
- AU first · VST3 included later
- Mono + linked stereo
- No instance tax
FEDUP 1.0$49ONE TIME. YOUR PLUGIN.
Free 1.x updates.
No subscription. No server.
HELP US BREAK IT
Put $49 in a bad room.
We’re looking for FOH engineers, worship teams, corporate audio crews, and venues willing to test ugly microphones in real systems.
Apply for early access ↗