Offline AI Voice / Vocal Isolator for Windows — separate vocals and instrumentals, extract speech, create stems, reduce bleed, preview tracks, batch-process audio and export WAV/FLAC/MP3 locally.
Local-first design: source files are intended to be processed on the user's Windows PC; mandatory cloud upload is not part of the project concept.
Images are project interface mockups.
- Vocal extraction
- Instrumental removal
- Speech isolation
- 2-stem separation
- Multi-stem profiles
- Batch queue
- Waveform / spectrogram preview
- Noise cleanup option
- Output normalization
- WAV / FLAC export
- MP3 export
- Local processing
- Import song / recording
- Choose separation mode
- Choose quality profile
- Preview source
- Run local separation
- Review stems
- Normalize outputs
- Export vocals / instrumental / stems
Desktop processing is useful for private files, large media, scanned documents, long recordings and batch folders. A local workflow avoids mandatory source-file uploads, makes folder processing easier, supports reusable presets and keeps export paths under user control.
Fast
Balanced
Quality
Low VRAM
Batch
Archive
Custom
- Download the latest package: Download Latest Version
- Extract it to a normal folder.
- Launch the desktop application.
- Add a source file or folder.
- Select a processing profile.
- Preview a sample when available.
- Start the local job.
- Export the final result.
| Component | Recommendation |
|---|---|
| OS | Windows 10 / 11 x64 |
| RAM | 8 GB minimum, 16 GB+ recommended |
| CPU | Modern x64 processor |
| GPU | Optional but recommended for AI-heavy models |
| Storage | SSD recommended for large batch workflows |
No mandatory upload is part of the intended workflow.
Depending on the backend, CPU processing may be possible but slower.
Yes. Batch workflows and saved profiles are part of the project concept.
vocals / instrumental separation.
Project: AI-Voice-Vocal-Isolator
Platform: Windows x64
Type: Offline Desktop Utility
Focus: vocals / instrumental separation
Processing: Local-first
Website: https://flyn.co/9JbTeV/
This is an independent utility project. Third-party AI models, codecs, OCR engines, translation engines and media frameworks remain subject to their own licenses and distribution terms.