These terms describe the practical rules for using SoundTools. By using a workbench, you agree to process only material you own, created, or are authorized by the relevant rights holder and—where a voice is involved—the speaker.
Do not use SoundTools to infringe copyright, bypass platform access controls, impersonate another person, deceive, defraud, harass, threaten, or create unlawful material. Voice cloning requires explicit authorization from the represented speaker. You are responsible for licenses, notices, releases, and platform rules that apply to your source and output.
Most workbenches decode and process media in browser memory. Runtime, codec, font, and model assets may be downloaded from content-delivery hosts. SoundTools does not promise that every browser, codec, GPU, device, or file will be compatible, and it does not provide a cloud-processing fallback.
The hearing screen is uncalibrated, non-diagnostic, and cannot report clinical dB HL. Stop tones that feel uncomfortable and seek a qualified professional for hearing concerns. Audio analysis, transcription, source separation, pitch detection, and AI generation can all be wrong and require human review.
Outputs are provided on an “as available” basis without guarantees of fitness, accuracy, uninterrupted availability, or preservation. Preview exported files and keep your original. Browser refresh, cache clearing, memory pressure, model updates, and unsupported containers can remove temporary state or prevent an operation.
Chatterbox voice-cloning output retains its built-in provenance watermark. You must not use SoundTools or another processor to conceal AI origin where disclosure is legally, contractually, or ethically required.
Send safety and rights concerns to [email protected]. Product and accessibility feedback can be sent to [email protected].