← Back

Prohibited content policy

Version in effect from: May 16, 2026

This policy is a detailed catalogue of content whose generation, entry into prompts or sharing in any way in the Yskra service operated by BADRED Sp. z o.o. is strictly prohibited. The policy applies both to content entered by the User and to outputs generated by AI models in response to the User's prompts.

§1. Zero tolerance — Category I (strictly prohibited)

The categories listed below result in immediate permanent Account blocking and, where the violation is a criminal offence, referral to law enforcement(Police, Prosecutor's Office, NASK Dyżurnet.pl, NCMEC):

  1. CSAM (Child Sexual Abuse Material)— any textual, visual or audio content depicting minors (under 18) in a sexual context, suggesting minor status of participants in sexual activity, "loli", "shota", "jailbait", "underage", "teen" in a sexual context (art. 200 and 202 Polish Penal Code).
  2. Content glorifying sexual violence — descriptions of rape, non-consensual sexual exploitation (non-consent) as a desired fantasy, incest.
  3. Terrorist material — instructions, propaganda, glorification of terrorism and genocide (Regulation (EU) 2021/784).
  4. Hate speech — incitement to hatred, violence or discrimination on grounds of race, nationality, religion, sexual orientation, gender identity, disability (art. 256, 257 Polish Penal Code).
  5. Totalitarian symbols — promoting totalitarian regimes (fascism, Nazism, communism) and use of their symbolism (art. 256 § 1 Polish Penal Code).
  6. Zoophilia, necrophilia, cannibalism — any content of this type (art. 202 § 3 Polish Penal Code).
  7. Non-consensual pornography— including "revenge porn", deepfakes of non-public persons without consent, intimate recordings of victims (Polish Act on Counteracting Domestic Violence, art. 191a Polish Penal Code).

§2. Category II (prohibited, account blocking)

  1. Impersonation of real persons — attempts to make AI pretend to be specific celebrities, politicians, family members, neighbours, ex-partners. See Content Removal Policy.
  2. Doxxing — sharing personal data of third parties (address, phone number, PESEL, card data) without their consent.
  3. Suicide or self-harm content in an encouraging or instructional manner (apart from emotional support, for which the AI is prepared — see the Terms of Service).
  4. Instructions for illegal activities — drug manufacture, explosive materials, firearms, hacking IT systems, tax fraud.
  5. Extreme graphic violence — detailed descriptions of torture, mutilation, gore in a context inconsistent with literary art (e.g. inciting real violence).
  6. Spam, scam, phishing — including attempts to extract passwords, BLIK codes, card data, cryptocurrency.
  7. Intellectual property infringements — generating content infringing copyrights, trademarks, related rights (see Copyright Policy).
  8. AI jailbreak attempts — deliberate attempts to bypass moderation to obtain any of the above categories.

§3. Permitted content in the NSFW context

The Yskra service enables flirty conversations and adult content with fictional AI characters. Permitted are:

  • conversations containing flirt, romance and emotional intimacy between adult AI characters and the User,
  • friendly, social and supportive conversations of the "digital pen-pal" type,
  • photos sent by characters in lifestyle / portrait / outfit / swimsuit / lingerie styles — comparable to mainstream fashion advertising (Calvin Klein, Victoria's Secret, Cosabella). No visible genitalia, nipples/areolas or sexual acts.

Yskra is not positioned or marketed as an adult-content or sex-chat platform. AI characters do not initiate or escalate conversations toward sexual content. Content in the Service is kept within mainstream-adult limits, comparable to popular social and messaging platforms.

Boundary: every theme has limits in the content of §1 and §2. If a conversation enters these categories, the AI model refuses to respond, and the moderation system may block the entire exchange.

§4. Detection and moderation

  1. The Service applies moderation layers: word filters, AI classifiers (detecting CSAM, suicide, terrorism, among others), pre-prompt restriction of the language model, and human control (manual review on reports).
  2. Attempts to bypass moderation are themselves a violation of the Guidelines and may lead to immediate Account blocking.
  3. If content from §1 is detected (CSAM, terrorism, gore involving real persons), the Service Provider:
    • blocks the account immediately and permanently,
    • archives evidence (in accordance with GDPR),
    • reports the matter to the relevant authorities: Police (998/112), NASK Dyżurnet.pl (dyzurnet.pl), and, where necessary, NCMEC (USA) and INHOPE.

§5. Reporting prohibited content

If you have noticed content violating this policy — report immediately:

§6. Final provisions

This policy forms an integral part of the Terms. The Service Provider reserves the right to update the list of prohibited categories where new threats arise or the law changes. Changes are announced on this page with an effective date.