SUNO watermark remover
A tool to remove watermarks from audio files.
Features
Core Audio Forensics & Analysis
- Spectral Fingerprinting Automated detection of the unique frequency pattern (often in the 18-24 kHz range) that identifies the specific AI model (Suno v3 vs. v4, etc.).
- Watermark Probability Score: A percentage confidence that a given file *contains* a hidden watermark, displayed before any processing.
- Segmented Detection: Analyzes the file in 1-second slices to find partial or intermittent watermarks within the track.
Advanced Watermark Removal Engine
- Inpainting Algorithm: Replaces the watermarked frequency bands with mathematically predicted "clean" audio data based on surrounding frequencies (classic spectral editing).
- Phase Cancellation Logic: Generates an inverse wave of the known watermark signature (requires a reference file of the watermark itself) to nullify it.
- Neural Denoising (AI-based): Uses a dedicated GAN to "hallucinate" the missing spectral data, ensuring smoother high-end frequencies than simple EQ filtering.
- Bit-Depth Re-Quantization: Breaks the hidden data by reducing the audio from 32-bit float to 24-bit or 16-bit PCM, then dithering (adds noise to mask the watermark hash).
Non-Destructive Workflow
- Project-Based Processing: Saves the original file untouched; all edits are applied to a virtual playback copy.
- Audition Preview: A/B toggle to compare the original vs. the "de-watermarked" output in real-time.
- Undo History: Full session logs so the user can revert any algorithmic action if the output sounds too "smeared."
Output & Metadata Management
- Lossless Export: Output as FLAC or WAV (recommended), with optional MP3 re-encoding at adjustable bitrates (192/256/320 kbps).
- Metadata Purge: Removes the "Produced with Suno" tags, XMP data, and embedder identifiers from ID3 tags.
- Fingerprint Alteration: Changes the file's acoustic fingerprint (acoustid) to prevent the AI platform from detecting the upload in the future—**though this is both unethical and a ToS violation**.
Conservation Tools (For Legitimate Use)
- DRM Stripping Check: Warns the user if the file is clearly copyrighted studio music (non-AI) and refuses to process it.
- Ethical Confirmation Prompt: A mandatory dialog box requiring a checkmark stating "I own the master rights to this audio or have explicit permission to modify it."
Critical Technical Realities (Warning)
- Not 100% Effective: Modern AI watermarks (like Suno’s) are designed to survive lossy compression, EQ changes, and noise addition. Removing them often audibly degrades the high-end (cymbals, air).
- Platform Detection: If you remove the watermark and upload to the same AI platform, their backend **will** re-detect the track via content fingerprinting and ban the account.
- Legal Risk: In the EU and US, circumventing a technological protection measure (even a watermark) is illegal under the DMCA and EUCD. This list does not condone piracy.
Summary Table
| Feature | Purpose | Risk Level |
|---|---|---|
| Spectral Fingerprinting | Identify the AI model | Legal to use (passive) |
| Neural Inpainting | Rebuild high-freq audio | Legal (processing your own file) |
| Reference Phase Cancellation | Null the watermark wave | Requires stolen watermark file |
| Metadata Purge | Strip ID3 tags | Legal (cleaning your own files) |
| Forced 16-bit Dithering | Destroy hash integrity | Legal (audio processing) |
| Content ID Evasion | Avoid platform detection | **Illegal & ToS violation** |
| DRM Override | Process non-AI studio audio | **Illegal** |
Download
Put downloadable files in projects/project-one/files/.
Screenshots
Put screenshots in projects/project-one/screenshots/.
Source code
Repository or source code link
License
Freeware / MIT License. You may use this software for personal, educational, or research purposes. Commercial use is prohibited without explicit permission from the author.