Uberduck AI Still Makes Raps—But Its Old Character Voices Are Gone

Uberduck is still operating as a creator platform for synthetic speech, private voices and AI-generated raps, but it is no longer the open-ended character-voice playground many early users remember. Uberduck’s explanation of the removed voices says it took down user-created models of characters and individuals in July 2023 because it lacked explicit consent and faced ethical and legal concerns.

The live service now presents a more conventional production proposition: users can start free, then pay for credits, private voice access, commercial rights and API access. Uberduck’s current plan comparison also lists AI-generated raps on the Creator and Pro tiers, confirming that music-oriented generation remains part of the product even though the old voice catalog does not.

What Uberduck AI is now

Uberduck is best understood as a collection of generative-media tools with voice at its core. Its public product categories include text-to-speech, voice cloning and music generation, while paid plans broaden the package with API and image-generation features. The service therefore overlaps several creator workflows rather than functioning solely as a text box that reads a script aloud.

The central voice workflow is straightforward in concept: create or obtain an authorized voice, provide text or musical material, generate an output and export it for a larger production. That can support narration, a temporary voice track, character dialogue for an original production or synthetic rap vocals. Output quality and suitability still have to be judged against the specific script, voice and intended distribution channel; the existence of a generation button is not evidence that every result will be publication-ready.

For businesses, the most consequential change is the shift in what supplies the product’s value. The attraction is now less about selecting an imitation of a recognizable celebrity or cartoon character and more about generating material with private or platform-provided voices under a defined plan. That is a narrower proposition, but it is also easier to evaluate as a production service.

Why the missing character voices matter

The removed library was not a routine interface redesign. Uberduck expressly connected the decision to the absence of explicit consent from the people or rights holders represented by those models. Users returning in search of a particular rapper, actor or animated character should therefore not assume that an old tutorial, screenshot or shared model name describes the present service.

This distinction also changes how Uberduck should be compared with alternatives. A catalog containing many recognizable names is not equivalent to a platform built around private clones and authorized voices, even if both products use the label “voice AI.” The relevant comparison is whether a service supports the user’s own approved material, provides the needed generation mode and grants a license appropriate to the final use.

There is no contradiction between the removal and Uberduck’s continued support for voice cloning. The former concerned models hosted without explicit consent; the latter can involve a user’s own voice or another voice for which the customer has permission. The technology remains available, but the responsibility for establishing authorization has become impossible to treat as a side issue.

Pricing is really a licensing decision

Uberduck’s displayed annual-billing rates put Starter at $2 per month, Creator at $5 and Pro at $30. The plans are differentiated not only by volume—1,000, 3,600 and 25,000 monthly credits respectively—but also by usage rights and production features. Prices and allowances can change, so a purchase decision should rely on the checkout terms shown at that moment rather than an older review.

Starter is explicitly non-commercial. That makes it suitable for exploration or private experiments, not automatically for monetized videos, client work, advertising or a commercial release. Creator and Pro list a commercial license, private voice access, API access and AI-generated raps; Pro adds a larger credit allocation and a stated 24-hour support response time.

Credits also prevent a simple price-per-month comparison. A creator should first estimate how often material will be regenerated after pronunciation, pacing or performance problems, because iterations consume capacity even when only the final take is published. Teams should additionally confirm whether the advertised commercial license covers their precise distribution, client and redistribution arrangements.

Consent and ownership remain the hard boundary

A technically possible clone is not necessarily an authorized one. Uberduck’s current terms for custom clones allow the company to verify that a customer is legally permitted to create the model; the terms also prohibit impersonation and infringement, say submitted content must be owned or lawfully usable, and state that private clone data will not be exposed to other users or used for other voice clones.

Those provisions make permission a prerequisite, not a cosmetic checkbox. For another person’s voice, a responsible production needs documented consent that covers the planned context, territory, duration and commercial use. Permission to record someone does not necessarily grant permission to build a reusable model, and permission for one campaign does not automatically authorize future scripts.

The terms further say users retain rights in submitted content while granting Uberduck a license to use, modify, reproduce, display and distribute content through the service. That language should be reviewed alongside any client confidentiality or talent agreement before sensitive recordings are uploaded. A company handling unreleased campaigns, regulated information or contractually restricted performances may need a separate enterprise agreement or a different workflow.

Where Uberduck fits—and where it does not

Uberduck is a plausible fit for creators who specifically want synthetic narration, a private voice, API-based voice generation or rap-oriented output in one account. Its low displayed entry prices reduce the cost of testing the workflow, while the separation between non-commercial and commercial plans makes one important limit visible before purchase.

It is a poor fit for anyone whose main objective is recreating the former catalog of famous or copyrighted characters. It is also not a substitute for securing performance rights, reviewing a script or conducting quality control. Businesses should retain the original consent record, the exact generated file, the final script and the plan terms that applied when the work was produced.

The practical verdict is therefore more specific than either “Uberduck is gone” or “nothing changed.” The service remains active and still includes voice and rap tools, but its operating premise has moved away from unapproved recognizable voices. For a current buyer, the decisive questions are whose voice will be used, whether that use is documented, which license covers the release and how many iterations the credit allowance can support.

Also read:

• Blog Link Building Still Works—But Its Old Shortcuts Now Count as Spam
• Gamma Imagine Makes Social Graphics Fast—but Exports Still Need Checking
• Character.AI Chats Feel Private—Its Policy Still Treats Them as Platform Data