Holly Herndon released her own voice as a governed instrument

Holly+, launched in 2021, turns audio into Holly Herndon's voice under a governance model her own studio describes but does not itemise.

Historical event

July 1, 2021

Site publication

September 18, 2026

What happened

In July 2021, the musician and composer Holly Herndon released Holly+, described on her studio's own project record as 'a machine learning instrument converting audio into Holly Herndon's voice with a shared governance model for artist identity'. The tool let anyone upload a recording and receive it back transformed into a rendering of Herndon's own voice profile — the same underlying capability that, in other hands, produces an unauthorised clone. The difference here is that the artist herself built and released the instrument, and framed its use around a decision-making structure rather than a simple licence.

What the documents say

Herndon and her collaborator Mat Dryhurst operate a studio archive that lists Holly+ as a 2021 'protocol' tagged protocol, voice and governance. The entry gives no further public breakdown of how the governance body votes or how proceeds are divided, and neither does the Holly+ product itself, whose only public-facing text, as retrieved on 16 September 2026, is the description 'Holly Herndon's digital twin'. That is a deliberately thin public record: the operative detail — who approves a derivative work, and on what terms — lives inside the tool's own interface rather than on a marketing page. A maker cannot verify a split or a vote threshold from outside; they can only verify that a governance layer exists at all, because the artist's own studio says so.

Why it matters for makers

Holly+ separates two things that unauthorised clones collapse into one: access to a convincing voice model, and permission to release what you make with it. Any voice-conversion tool can produce the audio; what Holly+ adds is a checkpoint an uploader has to pass before a track goes anywhere public. For a producer, that checkpoint is the entire difference between working with a licensed instrument and generating a sound-alike nobody agreed to. The mechanism worth naming is consent-at-source: the artist decides, through whatever process she has built, rather than a platform's takedown process deciding after the fact.

What to check before you use it

Before treating any 'artist-approved' voice tool as cleared, a maker should establish, in writing if possible, three things the public pages above do not settle: whether the specific use — commercial release, sync, remix — needs separate sign-off beyond upload; what happens to royalties if the track earns money; and whether approval is documented anywhere the maker can point to later if a rights-holder disputes it. This is editorial caution rather than something the cited pages state outright; the studio record and the product page confirm a governance layer exists, not its mechanics.

• Does the tool document who approves a release, or only that approval happens?
• Is there a written record of consent a maker keeps in their own files?
• Does the arrangement cover streaming only, or also sync, remix and resale?

Holly+ is cited across the industry as a reference case for artist-initiated voice licensing precisely because it inverts the usual sequence: consent first, generation second. That reputation rests on the existence of a governance structure, which the artist's own materials confirm, not on public disclosure of its terms.

Sources & reading trail

The artists' own studio site dates Holly+ to 2021 and describes it as a governance-tagged voice protocol.

Source published: Not established · Retrieved: 16 September 2026

The product's own site describes it as 'Holly Herndon's digital twin', confirming the artist operates it directly.

Source published: Not established · Retrieved: 16 September 2026

Papers, terms and official documents establish the record; the maker reading and the checks are Signal to Song editorial analysis. This retrospective draft does not imply the site published on the event date.

Continue reading

• A voice clone’s consent has to satisfy a statute and a platform
• Grimes offered her voice as a licensed model with a royalty split
• Browse the complete the archive

Published September 18, 2026, not on the date of the event described.