Frequently asked questions.

Yes — completely free with no sign-up, no watermarks and no usage limits. Because all processing happens on your own device, there are no server costs to pass on to you.

No. There are no accounts, no email confirmations and no passwords. Open the site and start working — your browser is all you need.

Anything your browser can decode: MP3, WAV, FLAC, OGG, M4A/AAC, OPUS and WebM. Exports are rendered to standard 16-bit stereo WAV files, which every editor and DJ app can open.

Songs up to around 5–6 minutes work best. Separation keeps six full stereo stems in memory, so very long files may hit your device's memory limit — a warning will appear if a file is too large.

Yes. The tools run in any modern browser — Chrome, Edge, Firefox or Safari — on desktop, Android and iOS. For long songs, mobile devices benefit from having fewer other tabs open.

No. The separation engine runs entirely inside your browser using the Web Audio API. Your audio files never leave your computer, which makes it safe for unreleased material.

We never receive your audio. The site uses Google Analytics to record anonymous usage statistics — pages visited, browser type and approximate region. See the Privacy Policy for the full picture.

A saved project is an ordinary file that lands in your Downloads folder — nothing is stored on a server. The tools also keep small preferences (like your last mixer state) in your browser's local storage, which never leaves your device.

Only for anonymous analytics (Google Analytics sets the _ga cookies). No advertising or tracking cookies are used. You can block or delete them in your browser settings without losing any functionality.

The on-device engine runs a spectral analysis — harmonic/percussive decomposition, frequency-band routing and stereo-position weighting — to route every sound into one of six stems: vocals, drums, bass, guitar, piano and other instruments.

Separation is an estimation, not a perfect reverse mix: sounds that share a frequency range (for example synths and vocals) can leak into neighbouring stems. Cleaner, better-produced sources give noticeably sharper results.

Two complete mix states you can flip between instantly. Set up mix A, switch to B, dial in different faders, pans, pitch and tempo — then compare the two with one click. Great for checking EQ-style decisions or before/after ideas.

A portable file that contains the audio itself plus every setting — faders, pans, pitch, tempo, loop region and both A/B slots. Save it anywhere, reopen it later from any device and continue exactly where you left off.

Yes. Solo the stem you want (press its S button) and use Export WAV — the render will contain only that stem. The same trick gives you a karaoke track (mute vocals) or an acapella (solo vocals).

The engine removes the parts of the mix concentrated in the vocal band and stereo centre, which works very well on typical pop and rock mixes. The music and vocals always sum back to the original song, so you can A/B the result against the source at any time.

The analysis reports the most likely key, scale and tempo together with a confidence level, and a tap-tempo button (press T) is there for instant cross-checks by ear. Highly rhythmic, tonal material scores best; heavily dissonant or free-form pieces can be ambiguous.

It makes Slowed + Reverb and Nightcore style remixes from any song — with independent speed, pitch, reverb amount and bass boost controls, plus MP3 and WAV download.

You can drag a region on the waveform, or use the keyboard: I sets the start point and O the end point at the playhead. Fade-in and fade-out ramps smooth the edges so your cut has no clicks.

To hear you, the browser must grant the page access to your microphone. The recording is processed and mixed locally, and the audio never leaves your device. Use headphones so the backing track doesn't bleed into the microphone.

The tools need JavaScript. Make sure it is enabled and that an aggressive ad-blocker or privacy extension isn't blocking scripts on this site, then reload the page.

Real-time pitch shifting uses AudioWorklet, which modern browsers only enable on secure (HTTPS) pages or localhost. Serve the site over HTTPS and every control is available; on plain HTTP everything else still works.

Six stereo stems plus analysis take real RAM. Close other tabs and apps, try a shorter file, or work on a device with more memory. On phones, processing a full 5-minute song is the practical upper limit.

Check that at least one stem is not muted and the master fader is up. If the download still fails, pause any download-blocking extension and try again — or save a project file and export from a fresh page load.

Write to info@tuti.lt with the page you were on, your browser and version, what you did and what happened. The more detail you include, the faster we can help.

The tool imposes no restrictions — but the copyright of the original recording belongs to its owners. Use separated stems only for purposes permitted by the rights holder (covers, practice, remixes you have clearance for, and so on).

You can use the tools themselves for any kind of work, commercial or not. What matters is the audio you feed in: processing and releasing copyrighted music commercially requires permission from its rights holders, which is between you and them.

No. We never see your audio and claim nothing over whatever you create with it. Your files, your mixes, your responsibility.