Frequently Asked Questions

No. The AIAPS watermark is embedded in the least significant bits of the audio samples at an imperceptible level (-144 dBFS) and is designed to remain inaudible under normal listening conditions while still being detectable by AIAPS verification tools.

No. The AIAPS Plugin operates locally on your machine. All analysis and watermark embedding is performed on-device. The only data transmitted externally is the fingerprint and metadata required for registry registration.

Yes. The perceptual fingerprint survives lossy compression including MP3, AAC, and OGG. The embedded watermark is preserved in lossless formats but may not survive lossy compression. The fingerprint serves as the durable identifier across all formats.

The NO_AI_TRAINING policy code indicates that this recording is designated as not authorized for AI training, dataset ingestion, voice cloning, style modeling, or synthetic generation without permission from the rights holder.

Traditional watermark systems embed a generic marker into audio. AIAPS watermarks are song-specific and encode identifying and policy information tied to the registry record. The watermark payload includes the AIAPS ID, policy code, and a checksum, embedded in the least significant bits of audio samples.

AIAPS v1.0 applies to finished music recordings and exported stereo masters. The plugin supports WAV, FLAC, and AIFF input formats. Fingerprint-based verification also works with MP3, AAC, and OGG. Additional format support may be added in future versions.

An AIAPS ID is a human-readable identifier assigned by the AIAPS registry in the format AIAPS-YYYY-NNNNNN, where YYYY is the registration year and NNNNNN is a zero-padded sequence number.

A tool may claim AIAPS v1.0 compatibility only if it generates an AIAPS-compliant fingerprint, embeds an AIAPS-compliant watermark, registers the recording in the AIAPS registry, supports the AIAPS verification workflow, and preserves required protocol fields and policy information. A tool that only writes metadata notices, only fingerprints audio, or only registers files without watermarking is not fully AIAPS-compliant.

AIAPS protection is performed using the AIAPS Plugin. The plugin fingerprints your recording, registers it in the AIAPS registry, and embeds an inaudible watermark. It is available as a standalone macOS app, AU plugin, and VST3 plugin.

No. AIAPS creates a public, verifiable record of the rights holder's stated terms: that a recording is not authorized for AI training. It establishes the declaration that platforms, AI systems, and legal frameworks can reference. Its effectiveness depends on adoption by those systems.