Capture a sample, then study the music.
SHAHI MUSIC PRODUCTION LAB v2.1
Import. Separate. Analyze. Create. Export.
A browser-based Shahi Music workspace for tempo and key changes, four-stem separation, music DNA analysis, song mapping, prompt creation, music-synced lyrics, vocal guidance and production reports for media you own or are authorized to use.
🔒 Local processing
01
Choose a file
Open the reference on YouTube
Some YouTube videos block embedded players. Shahi Music Lab v2.1.1 avoids that problem by capturing audio from the YouTube tab you explicitly share.
YOUTUBE REFERENCE STUDY
YouTube reference
—
Open the reference on YouTube, pause it at the point you want to study, return here and capture the YouTube tab audio. The app does not download the video.
Capture YouTube tab audio for Music DNA1) Open the song on YouTube. 2) Return to Shahi Music Lab and click Capture. 3) In Chrome/Edge choose the YouTube tab — not this Shahi tab — and enable Share tab audio. The sample stays in your browser and becomes the active analysis source.
Capturing…60s
Run lyric study after capture. Separate vocals first for best results.
Some YouTube videos refuse iframe embedding, and embedded players do not expose raw audio to webpage JavaScript. Shahi Music Lab therefore opens the reference in a normal YouTube tab and analyzes only the tab audio you explicitly share through your browser. Full source-lyric transcription is shown only when you confirm you own or are authorized to transcribe the recording; reference-only mode reports lyric structure without reproducing the source lyrics.
—
—
00:00
02
Tempo controls
Tempo
100%
New duration
—
Load a file to preview.
03
AI • Local
AI 4-stem separation
Separate the mix into production stems
Creates Vocals, Drums, Bass, Other and a combined Music / Instrumental WAV. HTDemucs runs in your browser; your song is not uploaded by this app.
Engine not loaded
First separation downloads an AI model of about 172 MB. WebGPU is used when available; otherwise it falls back to WebAssembly.
Best results
Use a clean stereo master. Separation quality depends on the original mix; reverb and tightly layered instruments can leave small artifacts.
Preparing AI engine…0%
The first run can take longer while the AI model downloads.
VocalsVoice isolated from the accompaniment
Music / InstrumentalDrums + bass + other music, with vocals removed
DrumsPercussion stem
BassLow-frequency bass stem
OtherHarmony, melody and remaining instruments
04
LOCAL ANALYSIS
Music DNA Analyzer
Analyze the instrumental or original mixEstimate BPM, key, tonal mood, energy, low-end weight, brightness, dynamics and an approximate song map. For the cleanest production analysis, separate stems first and analyze Music / Instrumental.
BPM—
Key—
Feel—
Mood—
Low end—
Brightness—
Dynamics—
Style estimate—
Time signature4/4 est.
Source—
Estimated song mapSection boundaries are snapped to musical phrase lengths and should be checked by ear.
05
SHAHI CREATOR
Create from the music
Lyric Fit Meter—
—
Estimated vocal range—Separate vocals first for range analysis.
Performance timeline
Analyze the music to build a section-by-section entry, pause and held-note guide.
06
Practice & production versions
07
Process & export
Original / authorized media only
This app improves export quality and workflow. It does not bypass copyright, fingerprinting, or platform recording-match systems.
Preparing…0%
The first run downloads the processing engine.
Processed file is ready
—
Pitch LockUses tempo processing to keep vocals near the original key.
Video SyncVideo timing is adjusted to match the new audio duration.
PrivacyYour media is processed in your browser, not uploaded by this app.