IsolateVocals Review: Vocal Separation, Costs and Exports

Short answer

IsolateVocals is a candidate for extracting a vocal and backing track from an authorized recording. Its simple two-track workflow suits practice and basic editing. If your job requires separate drums, bass and individual instruments, establish those requirements before selecting a voice-versus-music tool.

Evidence and review scope

We reviewed official product, pricing and privacy pages on November 14, 2026. We did not process audio, audition downloaded stems or purchase minutes. Ratings are withheld and links are direct. The listening exercise below is a proposed evaluation, not a benchmark result.

Inputs and output boundaries

The product page accepts common audio formats and MP4, M4V or MOV video, with a 20-minute ceiling. Audio can be up to 200 MB and video up to 1 GB. The documented output is two stereo WAV files at 24-bit/44.1 kHz. Its speech workflow separates voices from music, rather than promising general noise removal.

A high-resolution file format does not measure separation quality. Decide whether a little vocal residue is tolerable for rehearsal or unacceptable in a final edit. Evaluate the specific downstream use, not the attractiveness of a download specification.

Pricing and retention are separate

The pricing page offers 10 minutes for $1.99, 50 for $4.90 and 150 for $9.90, with usage counted to the second. The first song is free without an account; Google sign-in adds ten minutes. Purchased minutes do not expire. Its stated cash-refund window is seven days for an unused pack.

The homepage says files are available for 24 hours. Non-expiring minutes therefore do not mean permanent project storage. Download an accepted result while it remains available, and keep enough context to repeat the work if necessary.

Proposed listening test

Use an original recording or material you have permission to process. Choose a short section with a solo vocal, a dense chorus and a quiet transition. Write down whether you primarily need the voice or the instrumental; each can fail in a different way.

Listen to the original, then each output at comparable loudness. On the instrumental, check for remaining words, pumping and missing musical transients. On the vocal, check consonants, reverb tails and traces of cymbals or accompaniment. Repeat through headphones and your ordinary playback setup.

Import the files into your editor and align them with the source. Check the beginning and ending as well as the loudest section. Record whether the accepted stem works in the intended mix without distracting artifacts. A stem that sounds unusual alone may still serve a practice purpose, while one that sounds clean briefly may fail in a sparse arrangement.

Privacy and the exit step

The privacy policy describes private Cloudflare storage, processing through Replicate and temporary signed media links. The provider says uploads and separated tracks are not used for model training. We have not audited those controls or deletion behavior.

Keep signed links private and use non-confidential material for the first test. Export the source, accepted tracks and a short note describing the purpose. Processing a recording does not settle permission to publish a derivative project; establish your source permissions separately.

Verdict

Choose IsolateVocals for a small, well-defined separation test, especially when a one-time minute purchase suits your workload. For a deeper evaluation routine, read how to check AI-separated stems. Save the IsolateVocals profile, and follow the project export checklist before the file window closes.