SoundID VoiceAI vs Spotify and TikTok Clone Bans 2026

VoiceAI Transforms Your Take. It Is Not a Song Generator.

This page is the plugin plus DSP clone rules. Use the related AI voice-cloning legal-risks article for publicity statutes, Copyright Office guidance, and consent paperwork. Stay here for what SoundID VoiceAI actually does in a session, what Sonarworks claims about presets, and why Spotify and TikTok can still pull a release that sounds like a living artist.

Sonarworks markets it as a voice changer inside the DAW (VST3, AU, AAX). Listed hosts include Cubase, Logic Pro, Pro Tools, Reaper, Ableton Live, FL Studio, and Studio One. Capture on the track, process locally or in the cloud, print in place—no browser round-trip for the capture.^[1]

The 29 July 2026 C-pop post is explicit: VoiceAI transforms timbre, not performance. You sing the line in your language and phrasing; the model supplies the voice on top of that take.^[2] That is the opposite of a prompt-to-full-song generator. If you never recorded a performance, this plugin has nothing to map. Not legal advice.

• What you feed it Your recorded singing, hum, or beatbox. Output follows that input’s timing and phrasing.^[1]
• What you get A licensed factory or expansion model: demo vocal colour, Unison (up to eight layers), or voice-to-instrument (drums, guitar, violin).^[1]
• What you do not get A named-artist clone of a living star, a text-to-song master, or a DSP-safe impersonation certificate. Sonarworks’ royalty-free sentence is about their presets, not someone else’s identity.

Perpetual Local Processing vs Token Cloud

Two modes, not one silent subscription. Perpetual is a one-time licence for unlimited local CPU processing. Pay-as-you-go spends cloud tokens. Product page (checked September 2026): perpetual €99, expansion packs €29—re-check the live store; this recap is not a price lock.^[1]

Local processing never leaves the machine; cloud uses encrypted transfers. Perpetual still needs internet (offline use not supported) plus about 4 GB free RAM. OS floor: macOS 11+, Windows 10/11.^[1]

C-pop Pack: 29 July 2026, Ten Licensed Voices

On 29 July 2026 Sonarworks announced the C-pop Voices Expansion Pack: 10 voices (5 male, 5 female) for C-pop, Mandopop, Cantopop, and blended tracks that want Chinese vocal character. The blog: licensed, ethically sourced material; singers consented and were paid. Price there: €29, three devices, perpetual licence required for local use.^[2]

The live page lists C-pop beside Rock, Kids, Pop, and K-pop, with Meng as a C-pop demo. Token users get packs under current terms; perpetual owners buy a local licence per pack. Packs are 10 voices—no à-la-carte voices.^[1]

That pack is a licensed palette, not a famous-singer switch. “Chinese vocal character” is not “this named C-pop star.” A client request to “sound like [living artist]” leaves the preset licence and enters DSP impersonation rules. Not legal advice.

Royalty-Free Presets Are Not a Named-Artist Licence

Sonarworks’ FAQ: presets are “entirely free from copyright restrictions”; the company claims no rights in your content; you retain ownership for commercial and non-commercial use. Hero copy: royalty-free presets built with paid, consenting artists.^[1] Read that as: vendor-cleared models, vendor not claiming the bounce.

That sentence does not licence impersonating a named living artist on Spotify, TikTok, or any DSP. It does not override publicity, trademark, or impersonation rules. An “AI version of [star]” is not safe because it went through VoiceAI. Related legal-risks article: statutes. This page: licensed model versus someone else’s identity.

• Safer VoiceAI jobs Scratch vocals on a factory/C-pop/K-pop model; Unison doubles from your take; temp choir; hummed melody as an instrument. Credit the tool if a DSP asks.
• High-risk jobs A vocal clearly recognizable as a living artist; titling it their “AI version”; lookalike art plus soundalike lead. Omitting the name in metadata is not a defence if the voice is obvious.^[4]
• What this page will not do It will not re-litigate state publicity law. Keep that on the related 2027 legal-risks page. Here the question is: plugin licence versus DSP clone ban.

Spotify: Unauthorized Voice Clones Get Removed

On 25 September 2025 Spotify restated: vocal impersonation is allowed only when the impersonated artist has authorized the usage. The post names unauthorized AI clones, deepfakes, and other vocal replicas. Responsible production tools are not the target—identity, spam, and deception are.^[3]

Artist Help is the operational text: Spotify removes music that impersonates another artist’s voice without permission—AI clone or any other method. It applies even if you are not pretending to be the artist, even if you brand it an “AI version,” and even if the name is missing from metadata but the vocal is clearly that exact voice. Limited parody exceptions exist; do not assume yours is parody.^[4]

The artist (or someone for them) files a Publicity / Likeness claim. Platform Rules also bar deceptive impersonation and harmful synthetic media sold as authentic.^[5] A factory-model print of your take is not, by itself, that violation. Selling it as a named-star clone is. Not legal advice.

TikTok: Label Realistic AIGC. Do Not Impersonate Real People.

TikTok’s AIGC help covers images, video, and audio generated or modified by AI, including a real person’s altered image, voice, or words. Label realistic AIGC. “Significantly edited” includes AI speech and face-swap. Unlabeled realistic AIGC can be removed; falsely tagging unaltered audio as AI is also a Terms violation.^[6]

Banned even when labeled: fake crisis/authoritative sources; public figures in fake bullying or endorsement contexts; anyone under 18; adult private figures without permission. Integrity and Authenticity guidelines require realistic-AIGC labels and prohibit content that harmfully misleads or impersonates others. Report path: Misinformation → Deepfakes, synthetic media, and manipulated media.^[6][7]

A VoiceAI demo is not automatically allowed because the preset is royalty-free. Label realistic AI audio. Impersonation of a real person (private figure, under-18 likeness, fake celebrity endorsement) can be removed even with a sticker. Re-read live TikTok help the day you post. Not legal advice.

Release Checklist: Plugin Licence, Then DSP Rules

Treat VoiceAI as a performance mapper with a vendor licence, then apply Spotify and TikTok on top. A FAQ screenshot will not reopen a takedown. Keep dated dry take, preset name, and terms URLs.

1. Record a performance you control
   Do not feed a ripped commercial vocal and call the output original. Capture in the DAW; keep a dry backup beside the processed clip.
2. Stay on licensed factory or expansion models
   Use Sonarworks’ named presets (including C-pop after 29 July 2026). Do not prompt, title, or cover-art the file as a living artist.
3. Read Sonarworks as vendor terms, not DSP terms
   Royalty-free / “we claim no rights in your content” is the plugin contract. Re-check the live FAQ the day you bounce.^[1]
4. Before Spotify, apply the impersonation test
   Would a reasonable fan hear a specific living artist? If yes, you need that artist’s authorization—or do not upload. Metadata omission is not a defence.^[4]
5. Before TikTok, label realistic AI audio and skip impersonation
   Turn on creator AIGC label for realistic processed vocals. Do not post under-18 likenesses, unpermitted private figures, or fake endorsements.^[6]