Fadr

Fadr Basic is described as free for life, with unlimited stems, remixes and DJ sets. Each upload is split into vocals, melodies, drums and bass, and the Stems page also detects key and tempo, follows the chord progression and exports MIDI. Downloads on Basic are high-quality MP3; lossless WAV is a Plus feature.

The main free limit is length. The Plus page says Plus raises the song upload limit from 10 minutes to 1 hour, so a long DJ mix or a live recording will not fit on Basic. The stems guide accepts WAV, MP3, M4A, AAC, AIF and FLAC files under 200 MB each, and several files can be dropped at once.

The Basic split is four broad stems. Plus switches to the Pro Stems model and adds finer separation, which matters when only one part of a group is needed, such as a snare for a sample or backing vocals for a mashup:

• Vocals: separate lead and background vocal stems.
• Drums: kick, snare and the rest of the kit as separate files.
• Melodies: piano, acoustic and electric guitar, strings, wind and other melodies.
• Delivery: 32-bit WAV downloads, uploads up to 1 hour, concurrent stems and unlimited storage.

Every page carries a “50% Off Fadr Plus” banner, but the plan table and the pricing FAQ both state $10 a month or $100 a year, and no reduced amount is visible before signing in. Treat any discount as a promotion shown at checkout, not as the standing price, and check the total before paying.

The Plus page also includes a comparison table with prices of $19.99 and $29.99 a month. Those figures belong to other stem services in that table, not to Fadr. Billing is recurring until you cancel, and the terms say a cancellation takes effect at the end of the paid term.

Separating a song does not change who owns it. The terms require that anything you submit does not infringe copyright or other rights, and that you hold the licenses needed for it. Stems from a commercial release are fine for private practice, transcription or study, but releasing a remix, sample or instrumental normally needs permission from the rights holders.

The same terms give Pebble a perpetual, transferable and sublicensable license to contributions submitted to the site, a clause written broadly enough to raise questions for unreleased material. For your own unreleased masters, keep a copy of the original and check with Fadr how uploads are stored before relying on the web app.

The Stems Plugin runs on Mac as VST3 or AU and on Windows as VST3. It has three modes: Live Mode stems audio as it plays, Record Mode records a timeline and runs the best model, and Upload Mode lets you drag audio in and stems out. Separation happens in the cloud, then stems are mixed in the plugin window.

Stems made with the plugin are saved locally and are not lost if Plus is cancelled, but making new stems needs an active subscription. Siren and DrumGPT are separate Plus plugins that create playable instruments and drum kits from text prompts.

API use requires Fadr Plus. The only paid endpoint, Create Stem Task, costs US$0.05 per minute of audio and returns vocals, drums, bass, melodies and instrumental stems, MIDI for vocals, drums and bass, the chord progression, key and tempo. The detailed Plus stems are not listed for the API.

Each Plus renewal adds up to $10 of API credit, capped at $10, and usage beyond that accrues to a balance. The account is charged when the balance passes a billing threshold that starts at $10 and rises with use, or at renewal if the balance is above $1.

1. Record the file type, length and size before upload.
2. Upload the song and wait for stems, key, tempo and MIDI.
3. Download the vocal stem and the other three stems as MP3.
4. Import them into a DAW at the original tempo and compare with the session stems you already have.
5. Listen for vocal bleed in the instrumental and guitar bleed in the vocal, and check the detected key and tempo.

Accept an instrumental with no audible lead vocal in the chorus and a detected tempo within 1 BPM of the session. Reject missing sections, audible artefacts across whole bars or a key off by more than a semitone. No result is reported until the exercise is run.

To clean speech in a podcast or voice-over rather than split music, compare Adobe Podcast and Auphonic. To edit a recording by its transcript after separating a voice, see Descript.