Clumi AI is an online audio toolkit built around AI source separation. You upload a song, video or podcast clip, and the engine analyzes the waveform to pull out vocals, instrumentals and, in the higher tier, dozens of individual instruments. The workflow is entirely browser-based: drag a file in, wait for the analysis, then preview and download separate WAV tracks. This makes it useful for karaoke makers, remixers, producers, DJs and language or music learners who want isolated parts without buying desktop software.
Beyond separation, Clumi bundles a cleaning suite. The noise and reverb removers target specific problems — hiss, crowd noise, wind, static, mouth clicks, echo — and are aimed at podcasters, voice-over artists and videographers who need to salvage imperfect recordings. A set of utility tools rounds things off: key and BPM finder, audio-to-MIDI conversion, pitch changer, audio cutter, slowed-and-reverb generator and nightcore maker.
The product runs on a free tier plus a paid Pro tier. The free tier supports common audio and video formats with a capped file duration and size and a limited number of free uses; Pro Separation expands the stem count, raises the file-length and size limits and adds priority processing. Because the free tier is generous and requires no account, Clumi works well as a quick first pass before committing to a paid separation service.