VOICEVOX Model Terms

Version: 2026-10 (Last updated: October 10, 2026)

These Model Terms form part of the Speech Synthesis Model Terms and apply to users who configure "VOICEVOX" for a show on Lisnify.

1. Model and upstream terms

The conditions for using audio generated with VOICEVOX are set by the VOICEVOX software terms of use and, in addition, by the terms of use of each character (voice library). The Company offers only those characters whose terms the Company has confirmed permit commercial use in the Company's service model (audio generated by the Company's systems and distributed by users as a podcast) with attribution as the sole condition (Section 4).

2. Summary

VOICEVOX voices come with per-character terms. Lisnify only ships an allowlist of voices whose terms permit our use, and automatically appends the required credit line (for example "VOICEVOX:ずんだもん") to each episode description and to the show description. You must not use these voices for: political or religious activities, solicitation, endorsement or criticism; praising or attacking specific individuals or organizations (including states); defamation or other content that offends or harms third parties; adult content; information products or solicitation of investments and financial trading; spreading medical misinformation; deliberate fakes; adult-entertainment businesses or anti-social groups; encouraging crime; misrepresenting the author of the voice, claiming it as your own, or presenting it as official content; or machine learning and voice-changer use. See the model terms page for the full text.

3. User obligations (prohibited uses)

The prohibited uses in the character terms linked in Section 4 differ from character to character. So that every character can be used under the same conditions, the Company adopts the union of the prohibited uses of all offered characters as the prohibited uses of these Model Terms. You must not use the Service, in a show configured with VOICEVOX, for any of the following purposes.

1. Political or religious activities, solicitation for them, endorsement or criticism of a particular political or religious position, or other political or religious expression
2. Praising, attacking, or criticizing specific individuals or organizations (including states)
3. Defamation or other content that offends or harms third parties
4. Adult content (sexual content or excessively violent content)
5. Selling or promoting information products, or soliciting investments or financial trading
6. Spreading medical misinformation
7. Deliberate fakes that present false content as fact
8. Use in adult-entertainment businesses, use by or involvement with anti-social groups
9. Encouraging criminal activity
10. Misrepresenting the author of the voice (the character's rights holder), claiming generated audio as your own creation, or presenting it as official content of the rights holder
11. Using generated audio as training data for machine learning or as material for voice changers

The list above includes subjects that news shows tend to cover (politics, religion, and evaluations of specific individuals or organizations). Please do not register RSS feeds covering these subjects for a show configured with VOICEVOX. Users are responsible for the content they generate.

4. Offered characters (allowlist)

The Company offers only the following 22 characters, whose terms of use the Company has reviewed. The Company re-checks each character's terms periodically; additions and removals are handled as revisions of these Model Terms. The linked terms are in Japanese.

5. What the Company does

1. Automatic credits: The Company automatically appends, at the end of each generated episode's description, one credit line per character used in that episode (the string in the "Credit" column of Section 4). It also appends the credit lines for the characters configured for the show at the end of the show description in the show's feed (podcast RSS). You must not remove these credits when providing generated content to third parties.
2. Voice only: The Company does not display or use character illustrations or character profiles in the Service. Character names are displayed for attribution.
3. Record of agreement: When you configure VOICEVOX for a show, the Company asks you to agree to these Model Terms and records the version and time of your agreement for that show.
4. Discontinuation: If the Company becomes unable to offer a character (for example, at the request of the rights holder or because of a change in the terms), the Company stops offering the character for new selection and notifies the owners of shows using it to change the voice before the next episode is generated (Model Terms, Section 4, Item 4).
5. Changes: Changes to these Model Terms (including adding or removing characters and changing the prohibited uses) follow the procedure in Section 4 of the Model Terms.

Revision history