Acceptable Use Policy

1. Scope

This Policy covers every kind of content on Wisaven:

• Text: prompts and instructions, rewrites, translations, summaries, product copy, documents, slide decks, scripts and lyrics.
• Images: AI-generated and AI-edited images, posters, logos, portraits and product photos.
• Video: AI-generated and AI-edited videos, talking photos, lip-sync videos and dubbed videos.
• Audio: synthetic speech, cloned and designed voices, audiobooks, voiceovers, songs, instrumentals and processed audio.
• Uploads: documents, images, recordings, voice samples and videos you upload, and web pages you add as sources.
• Public content: songs, images, videos and voice clips you publish to the gallery, and your profile.

2. Content standards and prohibited categories

Wisaven prohibits generating, uploading or publishing these six categories of content:

• Pornography and sexually explicit (NSFW) content — of any kind. Wisaven does not allow adult content.
• Violence and gore — content that glorifies, incites or graphically depicts violence, self-harm or cruelty.
• Hate speech — content that attacks or demeans people for their race, ethnicity, national origin, religion, caste, sex, gender identity, sexual orientation, disability, age or serious illness.
• Child-unsafe content — any sexualisation of minors (CSAM), and any content that endangers children.
• Deepfakes and impersonation — content that impersonates a real person without their consent, including cloning or imitating their voice, swapping, animating or lip-syncing their face, or presenting fabricated words, actions or documents as real.
• Copyright and trademark infringement — reproducing existing books, articles, songs, recordings, images, video or artwork, or using brands and trademarks, without permission.

Zero tolerance. The following are actioned immediately and may be reported to the authorities: child sexual abuse material or any sexual content involving minors; content that promotes or facilitates terrorism, violent extremism or mass violence; instructions for weapons of mass destruction; incitement to genocide or ethnic hatred; and non-consensual sexual deepfakes of real people, in image, video or voice.

Allowed only in context. Creative and informative work may deal with conflict, crime, war or grief, but must not glorify real attacks, incite violence against anyone, or give instructions for causing harm. Content may express political opinions, but must not spread disinformation about elections, public health or emergencies, or impersonate candidates or officials. Strong language is allowed in creative work, but not as targeted harassment of a real person.

Also prohibited: presenting misinformation or disproven claims as fact; forged or fraudulent documents, IDs, certificates, receipts or reviews; impersonating real people, brands or organisations, including passing work off as their official output; cloning a voice that is not yours without the speaker’s express permission; using anyone’s face for face swap, talking photos or lip sync without their consent; uploading documents, recordings, images or videos you have no right to use; passing AI-generated content off as wholly human-made where that would mislead; bulk-generating spam or using output for deceptive commercial purposes; and jailbreak prompts or other attempts to get around our safety measures.

3. Voices, faces and uploads

Some tools work on a real person’s voice or likeness. For these, the following rules apply in addition to the categories above:

• Voice cloning: clone only your own voice, or the voice of someone who has given you their express permission. You confirm this before each voice is cloned, and we keep a record of that confirmation.
• No impersonation: never use a cloned or designed voice, or a swapped, animated or lip-synced face, to impersonate a real person without their consent, to make people believe someone said or did something they did not, for fraud or scams, or to get past voice-based identity or security checks.
• Uploads of people: tools that work on a person’s photo, video or recording ask you to confirm that it is yours to use, or that the person agreed. Do not use a public figure’s face or voice.
• Private results: audio spoken with a cloned voice, and results made from someone’s photo, video or recording by tools that say so, stay private to you — you can download them, but they cannot be published in the gallery.
• Removal: we may remove any voice and any audio, image or video made with it or from an upload, and restrict these tools, when we receive a credible complaint or believe these rules have been broken. If someone has cloned your voice or used your face without your consent, report it to [email protected].

4. Content moderation

We run layered moderation: automated screening, human review and periodic checks.

• Before generation: the text of every request — prompts, scripts, lyrics, descriptions and chat messages — is checked against our blocklist of prohibited terms, which our trust & safety team maintains and keeps updated, and screened by Waffo Prompt Sift, a content-safety service for AI generation. Requests that fail are not generated and not charged.
• After generation, before delivery: generated images and video frames are checked by a vision-model safety classifier. A result that fails is withheld and its credits are refunded.
• Before publishing: everything is private until its owner publishes it, and some results can never be published (see “Voices, faces and uploads”). The text of a work is checked against the blocklist again when it is published. Moderators can review, hide or remove any published work, and anything a moderator has turned down cannot be published again without their review.
• Reports: anyone signed in can report a published work. Each person’s report counts once; when three people report the same work, it is hidden until a moderator has looked at it.
• Usage patterns: rate limits, and monitoring for unusual automated use.
• Human review: submitted and reported content goes to a review queue handled by our trust & safety team, who approve, turn down, keep or take it down. Reviewers are trained on this Policy and bound by confidentiality.
• Periodic checks: at least quarterly, we sample published work and moderation decisions, and update the blocklist and filters based on what we find.

Severity levels: L1 Critical (child safety, terrorism) — blocked automatically where possible and confirmed by a person; immediate permanent ban and report to the authorities. L2 High (other severe violations, including non-consensual deepfakes) — reviewed first; content removed and account suspended. L3 Medium (general violations) — human review; warning and removal. L4 Low (minor violations) — reviewed when reported; warning and edit required.

We keep records of automated decisions and human reviews.

5. Reporting

Please report content or behaviour that breaks this Policy — especially anything involving child safety or terrorism — and AI impersonation of you, including a cloned voice or your face used without your consent, or infringement of your rights.

Reporter identities are kept confidential. Every report is reviewed impartially. Malicious or knowingly false reports are recorded and may themselves be actioned.

• On any published song, image, video or voice clip: the Report button.
• Email: [email protected] — include the link and what is wrong.
• Copyright claims: [email protected] (see wisaven.com/legal/copyright).

6. Response times

• L1 Critical: acted on within 24 hours.
• L2 High: first response within 24 hours; resolved within 3 business days.
• L3 Medium: first response within 3 business days; resolved within 7 business days.
• L4 Low: first response within 5 business days; resolved within 15 business days.

In-product reports are acknowledged immediately; email reports are acknowledged with our first response. If a case needs longer, we tell the reporter why. The workflow is: received → classified (L1–L4) → reviewed (automated and human) → decision → action → notification.

7. Enforcement

• Warning — a written notice requiring removal or an edit. For a first minor violation.
• Content removal — the content, or a voice, is taken down. For confirmed L3–L4 violations.
• Feature restriction — publishing, voice cloning or generation paused. For repeated minor violations or a single L3.
• Account suspension — temporary, pending review. For L2 or repeated violations.
• Permanent ban — access ends and unused credits are forfeited without refund. For L1 or repeated L2.
• Legal action — reporting to the authorities and cooperating with investigations. For criminal L1 violations.

8. Appeals

If you think we got it wrong, you can appeal within 30 days of our notice: email [email protected] with the subject “Appeal”, your account email and the decision you are appealing. Appeals are reviewed within 10 business days by someone other than the person who made the original decision, and we tell you the outcome.

9. Changes to this Policy

We notify material changes at least 15 days in advance. Continuing to use Wisaven after a change takes effect means you accept the updated Policy.

10. Contact

• Reports, safety and appeals: [email protected]
• Copyright: [email protected]
• Company: Shenzhen Xin Wei Xin Technology Co., Ltd. (深圳心为心科技有限公司)
• Address: Rm 315, No. 2-46 Minsheng Avenue, Shangcun Community, Gongming Subdistrict, Guangming District, Shenzhen, Guangdong 518107, China