Hi Product Hunt 👋 I'm Kaz, an indie iOS developer from Japan. I built VoiceKit because I kept re-recording videos over a single "um." Existing tools do this well — but they upload your audio to a server, and they're subscriptions. For a meeting recording or a lecture, that felt wrong. So VoiceKit does everything on the device. Transcription, filler detection, subtitles, and even summarization run on your iPhone. Not one byte of your audio leaves it. One thing I learned the hard way: Whisper doesn't transcribe hesitations. "Um," "uh," "えー" — it silently drops them, on every model size. So a dictionary-based approach can't work. VoiceKit instead compares voice activity against Whisper's word timeline: where there's speech but no word, that's a filler. It's a bit backwards, but it works. It's also honest about its limits. Automatic detection won't catch everything — a quiet "um" that runs into the next word is genuinely hard. So there's a manual mode: play, mark IN and OUT, done. Automation does the prep; your ears do the finish. One-time purchase. No subscription, no account, no tracking. Happy to answer anything — especially about the on-device side.

VoiceKit

Remove filler words from video — entirely on your iPhone

Remove filler words from video — entirely on your iPhone