Free tools
Free audio tools
Free, with none of the usual catches: no account, no email, no watermark, no ads. Five run entirely in your browser, so your file never leaves your device. The vocal remover runs on the Euphona Engine, and its page says so.
LUFS and true peak meter
Integrated LUFS, true peak, peak-to-loudness ratio and loudness range to ITU-R BS.1770-5, with your track set against the streaming platforms’ targets and who publishes each.
Open the LUFS and true peak meter →BPM & key finder
Tempo and key, with the half-time and relative-key alternatives shown wherever the answer is genuinely ambiguous rather than hidden behind a single label.
Open the BPM & key finder →Spectrum analyzer
The whole file's average spectrum in third-octave bands, with the tilt and centroid figures that make two mixes comparable: a measurement, not a light show.
Open the spectrum analyzer →Stereo width analyzer
Phase correlation, width and what actually survives a mono fold-down: overall and in five bands, which is where width problems live.
Open the stereo width analyzer →Clipping & noise detector
One report: clipped runs with timestamps, true peak against sample peak, the noise floor, mains hum, DC offset. Detection is free; repair is what restoration is for.
Open the clipping & noise detector →Vocal remover
Upload a song and hear the instrumental and the vocal, separated by the Euphona engine and played level-matched against the original. The one tool here that uploads, and deletes what it took within 24 hours.
Open the vocal remover →
What the measurements follow
Loudness is measured to ITU-R BS.1770-5, the algorithm EBU R 128 is built on and the one Spotify names for its loudness normalization. That means integrated loudness in LUFS, true peak in dBTP measured on a 4× oversampled signal, and loudness range in LU. These are defined quantities, not opinions: a compliant meter given the same file returns the same numbers, which is why a reading here can be checked against any other conformant tool.
Tempo and key detection are estimates rather than measurements, and the tool says so where the answer is genuinely ambiguous: a track at 140 BPM and the same track read as 70 BPM are both defensible, as are a major key and its relative minor. Showing the alternative is more honest than picking one and presenting it as certain.
Which of these sees your file
The loudness checker and the BPM and key finder decode and analyze your audio in the browser, with the Web Audio API on your own machine. Nothing is uploaded, so there is nothing for us to store, log or delete, and once the page has loaded it works with the network disconnected.
The vocal remover is the exception, and it's labeled as one on its own page. Separating a mix into stems needs a model that can't run usefully in a browser, so that file is uploaded, processed on the Euphona engine, and deleted within 24 hours. It's the only tool here that leaves your device.