Voice Cloning

What Is Voice Cloning?

Voice cloning is the synthesis of speech that imitates a specific person's voice from a sample. In a companion app the language model writes the reply and a separate speech model reads it aloud, so a custom voice changes the speech step, not the personality (see how AI girlfriend apps work).

Voice Cloning in Named Apps

• CrushOn.AI: its wiki describes custom voice cloning from about 30 seconds of audio, for private use only, with a consent checkbox (we did not capture its wording). Custom voice slots are 1 on Free, 3 on Standard, 5 on Premium, 10 on Luxe, 20 on Elite and 30 on Imperial. Voice messages you record are deleted after 24 hours.
• Character.AI: characters can speak replies with a pre-made or user-created voice, and the Creator Guide describes a 3 to 15 second voice sample. Character Voice is free, and its privacy policy lists voice recordings among data collected.
• SpicyChat: Custom Voices are created from uploaded audio samples on the top I'm All In tier.
• Nastia AI: the install page lists "Custom Voices - clone" on its paid plan.
• Kindroid: custom voices by voice design or from samples, with V2, V3 and V4 voice models. Its help center mentions an unnamed voice provider.
• Nomi: Custom Voices with unlimited usage, plus an ElevenLabs integration that uses your own ElevenLabs allowance. How Custom Voices are made is not documented.

For SpicyChat, Nastia AI, Kindroid and Nomi we found no published consent rule. Compare costs in our ranking of AI girlfriends with voice.

Consent Rules and Laws

Each rule below applies only where it says.

• Microsoft: its personal voice feature requires "explicit consent from the user" and a recorded statement acknowledging that the voice will be created and used.
• ElevenLabs: "Voice-captcha technology is used to verify that Professional Voice Clones are created from your own voice samples."
• United States: the FCC ruled that calls made with AI-generated voices are "artificial" under the Telephone Consumer Protection Act, which covers robocalls, not companion chat.
• European Union: Article 50(2) of the AI Act requires providers of systems that generate synthetic audio to mark outputs as artificially generated in a machine-readable format. Providers whose generative systems were already on the market before August 2, 2026 have until December 2, 2026.

The companion-chatbot laws we track, such as California's SB 243 and New York's Article 47, deal with AI disclosure and crisis protocols. The summaries we read do not mention voice cloning.

Privacy of Voice Samples

A voice sample is personal data, so read the app's policy before uploading one. See whether apps store your chats and the AI companion privacy section. More terms are in the glossary.