Acceptable Use Policy

Effective:30 October 2026. Operator:Wu Wenlin (巫文林), an individual developer in China. This policy covers Tiny Desk inputs, photos, audio, video, accounts and sharing, and forms part of the Terms. Report abuse or appeal to support@tinydesk.pro.

1. Six prohibited content categories

1. Sexual/NSFW: pornography, sexualized nudity, sexual exploitation, non-consensual intimate imagery and nudification. No adult-content exception is offered.
2. Violence/gore: realistic extreme gore, abuse, threats, terrorist promotion, incitement to real harm and weapons-of-mass-destruction instructions.
3. Hate speech: dehumanization, persecution or discriminatory violence targeting race, ethnicity, religion, nationality, gender, sexuality, disability or similar characteristics.
4. Child unsafe content/CSAM: sexualization, grooming, exploitation or abuse involving minors, including synthetic, fictional or apparently underage depictions. Do not upload children's portraits.
5. Deepfakes/impersonation: real-person face swapping, nudification, fabricated intimate interactions and invented statements or actions attributed to a real person. Do not impersonate people, public figures, brands or institutions. Consent or an AI label does not exempt prohibited uses.
6. Copyright/trademark infringement: unlicensed protected music, footage, characters, marks or likenesses, and false claims of rights-holder or model-provider endorsement.

Fraud, malware, harassment, privacy violations, spam and bypassing safety checks or bans are also prohibited. Upload only your own adult material or expressly authorized material. Ownership alone does not authorize prohibited uses.

2. Platform moderation

Generation/upload availability and verified limits are shown in the live interface. Input safety screening is mandatory before model submission.

Before a model call, the driving text—server template, role assignments, model instructions and applicable user direction—must pass Waffo Prompt Sift / Prompt Screening API. Only an explicitallow permits continuation.block rejects;review, timeout, errors, missing credentials and malformed responses defer generation. Photos and image URLs are not sent to that text endpoint. Missing credentials keep generation disabled.

APIMart generation requests enablensfw_check=true; its checks do not replace ours. Prompt approval is not output approval. Output safety still requires provider capabilities and operator investigation of reports; we do not claim a completed independent automatic output scanner. The operator manually investigates reports concerning delivered content and may restrict access or features first. We do not claim unattended automatic approval or an external 24/7 review team.

Necessary safety records contain decisions, time, request identifiers and prompt hashes, not private keys, full photo links or raw prompts. Rules and handling records are reviewed weekly and after safety incidents. Unreviewed raw output must not enter a public gallery; no public user gallery is currently offered.

3. Reporting and response times

Email support@tinydesk.pro, subject“Content report”, with the page or task identifier, category, facts and minimal necessary evidence. Never send suspected CSAM files or redistribute unlawful material. Reporter details are shared only as necessary for handling or lawful authority requests.

• Child-safety, imminent harm or serious illegality: confirmed prohibited requests/access are restricted immediately when identified; reports receive priority human assessment within 24 hours. Authorities are involved where required.
• Other high-risk reports: initial response within 1 business day, normally a decision within 3 business days.
• General reports: acknowledgment within 2 business days, normally resolved within 7 business days. We explain evidence requests or extensions.

These are handling targets for an individually operated service, not a live emergency response service. Contact local emergency services for immediate danger.

4. Enforcement and credits

Minor first breaches may receive a warning and correction request. Confirmed violating material is refused or removed; model, upload or sharing access may be restricted. Serious or repeated abuse can cause immediate suspension followed by termination after review. The operator executes measures through administration tools and infrastructure and explains reasons unless law or safety prevents disclosure. Necessary evidence may be preserved and serious suspected crimes reported lawfully.

We do not automatically confiscate all legitimately purchased balance. Fraudulently obtained or abusive credits may be revoked; other unused prepaid value is handled under the Terms and mandatory law. Bans, refunds and evidence preservation are reconciled separately to avoid duplicate deductions.

5. Appeals and changes

Within 30 days of a decision, email support from your account address, subject“Content appeal”, with identifiers and reasons. The operator rechecks the facts and records, normally responds within 10 business days and corrects errors, restoring access or allowance where appropriate. We do not claim a separate independent review team. Material changes normally receive 14 days' notice on the website or by account email; urgent legal/safety changes may apply immediately.