Mini mixer
On Home, expand Extract layers to solo stems and Play all in sync before export.
Processing in background…
Where to next
Stems
Unmix your mix
Vocals, drums, bass, and more — each on its own lane, ready to remix, rehearse, or polish.
Listen
Hear it before you ship it
Waveforms, layers, and quick A/B so you know it’s right — without leaving the page.
Let the music out.
Upload a track — get a fast five-stem split: vocals, drums, bass, synth, and other — plus sub / mid bass from DSP.
+
Drop an audio file or click to upload
MP3, WAV, FLAC, M4A, OGG · up to 100 MB · 5 credits per job
No file selected
0%
Average stem generation takes 2-3 minutes to achieve the highest quality.
Stems
Unmix your mix
Vocals, drums, bass, and more — each on its own lane, ready to remix, rehearse, or polish.
Listen
Hear it before you ship it
Waveforms, layers, and quick A/B so you know it’s right — without leaving the page.
Now playing
—
Project label
Optional tag for grouping tracks — set it from Library → Edit → Set project…
- Original file
- —
- Stems
- —
- Added
- —
- Upload size
- —
Stem mix
Chord
—
MIDI & song analysis
Get separate MIDI files (vocals, bass, drums, melody from synth + other), plus estimated musical key, BPM, and a chord progression from your stems. Same credit cost as drum / vocal layers (approximate transcription — audition in your DAW).
Typical run: 1–3 minutes depending on track length.
Key and tempo appear to the right of the layer tabs; the live chord sits under the waveform. Use MP3 / WAV / MIDI on each stem row below.
Vocal layers
Split the vocal stem into dry lead, backing, doubles, and room — 5 credits, same as drum layers. One automatic pass (no presets).
Typical run: 2–3 minutes for highest quality.
Mini mixer: Play all mixes the unmuted layers in sync (like drum layers). Together they approximate the full vocal stem. Solo / Unsolo on the same button.
Drum layers
Split the drum stem into kick, snare/clap, hi-hats & cymbals, and toms & percussion (5 credits — same as one stem job).
Typical run: 2–3 minutes for highest quality.
Mini mixer: Play all mixes unmuted drum layers. Solo / Unsolo on the same button. Mute is the switch; red ✕ on the lane when muted.
Trimmed samples (this track)
Use In / Out at the playhead on each drum layer, then Save sample. Download WAV below.
No saved samples yet.
Other — deeper split
Optional: split the other stem (beds / FX) into harmonic vs FX — runs only on that layer after drums, bass, and synth are separated. Same credit cost as drum / vocal layers.
Typical run: 2–3 minutes for highest quality.
Mini mixer: Play all mixes unmuted harmonic + FX layers. Solo / Unsolo on the same button. Mute is the switch; red ✕ on the lane when muted.
Explore
Browse your stem library like a catalog — open tracks, preview layers, and jump straight into separation workflows.
0 ready tracks ·
Jump back in
Recently in your library
Browse by stem
Open the Library filtered by layer — use the stem tabs and search bar there.
Tips & workflows
Projects & covers
Group jobs with a project label from Library → Edit — great for albums and A/B stems.
Studio
Studio is the DAW-style stem workspace. Vocal styles (RVC) lives on its own page — open it from the sidebar when you need a style transform.
Studio
Project and remix live in the left strip next to the main menu. Mix stems here — Replace on a row for one layer only.
Loading stems…
♪
—
100%
Master
0−6−12−24−48
Vocals
No stem
0−6−12−24−48
Drums
No stem
0−6−12−24−48
Bass
No stem
0−6−12−24−48
Synth
No stem
0−6−12−24−48
Other
No stem
Vocal styles
RVC style transform on your separated vocals stem — separate from the Studio workspace. Pick a track and a preset; processing runs on the server.
Same completed jobs as in your Library.
Vocal style transform
RVC
Backend: vocal_transform.py →
vocals_{style}.wav
Cost: 5 credits per vocal transform (same as one stem job).
?
—
@handle
0 jobs
0 credits
—
Featured
No completed tracks yet — upload on Create.
Your library
Account & billing
Stem separation is 5 credits per job (about $0.25 per split). Buy packs when you need them — no subscription, no automatic renewal.
Model
Pay as you go
Credits
0
Recurring billing
None
Credit packs
Pick a pack to checkout securely with Stripe. Your balance updates after payment.
100
credits
$4.99
~20 splits · ~$0.25 each (5 credits)
Popular
300
credits
$14.99
~60 splits · ~$0.25 each (5 credits)
1000
credits
$49.99
~200 splits · ~$0.25 each (5 credits)
One-time charges only. No “cancel plan” needed — there is no plan to cancel.
Account details
Display name
—
Email
—
Google
—
User ID
—
Password change and two-factor authentication will be available in a future update. For billing questions, contact hello@melodinai.com.
Admin
Users, uploads, and stem processing (visible only to site administrators).
Remote GPU (RunPod Serverless) and custom worker URLs. Values are stored in
config/api_secrets.local.php on this server (not in git). Shell
environment variables such as MELODIN_RUNPOD_API_KEY still override
these when set (e.g. in Plesk PHP-FPM).
—
Outgoing mail (welcome emails, etc.). Credentials are stored only in
config/mail_secrets.local.php on the server — not in app source.
Plesk can show “Sent” or queue messages while delivery to Gmail stays Deferred if this VPS
cannot use outbound port 25 — the admin panel will warn when that probe fails. To use
smtp.ionos.com (587, TLS), set it in
config/mail.local.php and save the SMTP password from the
IONOS email account (Customer Panel / webmail); it is often not the same
as a password only changed inside Plesk, which causes “Could not authenticate”.
—
Library
Your processed tracks, workspace, and stem picks — search by name or project below.
Stems
No matching tracks.
Samples
One-shots and loops you trimmed from stems — filter by category, assign labels on Create, and organize here.
No samples yet.