Heartbits AISubscribe

HeartBits Prohibited Content Policy

Effective Date: July 13, 2026
Last Updated: July 13, 2026

1. Purpose

To protect the safety of HeartBits (the "Platform") and the lawful rights of users and third parties, the Platform maintains a prohibited-term and prompt-filtering library and reviews user prompts, character settings, uploaded materials, and AI-generated results for safety risks.

This Policy applies to all content submitted, generated, edited, published, shared, downloaded, or distributed through HeartBits.

2. Core Principles

The Platform manages prohibited content according to the following principles:

  1. Legal Compliance: The Platform may not be used to create or distribute content that violates applicable law.
  2. Child Safety: Strict restrictions apply to sexual, violent, exploitative, grooming, or otherwise inappropriate content involving minors.
  3. Protection of Rights: Users may not infringe intellectual property, likeness, privacy, reputation, or other lawful rights.
  4. Prevention of Real-World Harm: Content that could directly facilitate violence, crime, fraud, self-harm, or other serious harm is prohibited.
  5. No Circumvention: Users may not evade safety controls through altered spellings, split words, homophones, encoding, coded language, prompt injection, or similar techniques.

3. Prohibited Content Categories

3.1 Illegal Activity and Criminal Assistance

Content that promotes, plans, carries out, or materially assists illegal activity is prohibited, including:

  • Criminal plans, operational steps, division of roles, or methods for avoiding detection;
  • Theft, robbery, kidnapping, extortion, smuggling, or unlawful detention;
  • Forged identification, documents, seals, currency, tickets, or certificates;
  • Money laundering, illegal fundraising, underground transactions, or concealment of criminal proceeds; and
  • Actionable instructions, tools, or resources intended to facilitate real-world crime.

Examples of high-risk prompts: "crime tutorial," "forge identification," "avoid police detection," "launder money," or "destroy evidence," including variations with the same intent.

3.2 Violence, Terrorism, and Extremism

The following content is prohibited:

  • Promotion, praise, recruitment, financing, or assistance for terrorism or violent extremism;
  • Credible threats of real-world violence against a person or group;
  • Graphic or glorified depictions of severe injury, torture, or dismemberment;
  • Attack planning or instructions intended to cause mass harm; and
  • Content intended to incite, celebrate, or encourage imitation of violent events.

Examples of high-risk prompts: "terrorist attack," "extremist propaganda," "violent threat," "graphic torture," or "attack plan," including variations with the same intent.

3.3 Weapons, Explosives, and Dangerous Materials

Content that directly assists the manufacture, modification, acquisition, concealment, or harmful use of weapons or dangerous materials is prohibited, including:

  • Making or modifying firearms, ammunition, explosives, or other offensive weapons;
  • Unlawful use of hazardous chemicals, biological agents, or radioactive materials;
  • Methods for evading security screening or unlawfully transporting dangerous materials; and
  • Converting ordinary items into weapons for the purpose of harming others.

Examples of high-risk prompts: "make an explosive," "modify a firearm," "dangerous chemical formula," or "bypass security screening," including variations with the same intent.

3.4 Sexually Explicit, Exploitative, and Non-Consensual Content

The following content is prohibited:

  • Explicit sexual content or content primarily intended for sexual arousal;
  • Forced, coerced, secretly recorded, drug-facilitated, or otherwise non-consensual sexual content;
  • Sexual exploitation, solicitation of commercial sexual services, or human trafficking;
  • Non-consensual synthetic intimate content; and
  • Use of a real person's face or identity in sexual, nude, or sexually suggestive content without authorization.

Examples of high-risk prompts: "explicit sexual content," "non-consensual sexual act," "secret intimate recording," "sexual deepfake," or "sexual services," including variations with the same intent.

3.5 Child Safety

Any sexualized, exploitative, grooming, abusive, or otherwise inappropriate content involving a person under 18 is strictly prohibited, including fictional, animated, synthetic, and role-play content.

Prohibited content includes:

  • Nudity, sexualization, sexual conduct, or sexual content involving minors;
  • Grooming, arranging inappropriate contact, or requesting intimate material from a minor;
  • Glorifying, normalizing, or instructing the exploitation or abuse of minors;
  • Depicting an adult as a minor for a sexual or sexually suggestive scenario; and
  • Severe violence, bullying, dangerous challenges, or criminal conduct involving minors.

Examples of high-risk prompts: Any prompt combining terms such as "child," "minor," "school student," or "young-looking" with sexualization, exploitation, grooming, or abuse, including variations with the same intent.

3.6 Hate, Discrimination, and Group-Based Attacks

Content that demeans, dehumanizes, excludes, promotes hatred against, or advocates violence toward people based on race, ethnicity, nationality, religion, sex, sexual orientation, disability, illness, age, or another protected characteristic is prohibited.

Examples of high-risk prompts: "eliminate a protected group," "a certain ethnicity is inferior," "expel a religious group," or "create hate propaganda," including variations with the same intent.

3.7 Harassment, Bullying, Threats, and Defamation

The following content is prohibited:

  • Persistent insults, threats, stalking, or malicious harassment directed at a specific person;
  • Encouraging coordinated harassment, doxxing, or mob attacks;
  • Fabricating damaging factual claims and presenting them as true;
  • Content intended to humiliate, coerce, retaliate against, or manipulate another person; and
  • Non-consensual synthetic content intended to humiliate or harm a real person.

Examples of high-risk prompts: "dox this person," "threaten someone," "expose private information," "invent a scandal," or "organize online harassment," including variations with the same intent.

3.8 Self-Harm, Suicide, and Dangerous Challenges

Content that encourages, glorifies, instructs, or materially assists self-harm, suicide, or other conduct likely to cause serious injury is prohibited, including:

  • Methods, steps, or tool recommendations for self-injury or suicide;
  • Encouraging another person to self-harm, die by suicide, or participate in a dangerous challenge;
  • Romanticizing self-harm or suicide as admirable or worthy of imitation; and
  • Manipulative or harmful content directed at a vulnerable person.

Supportive, preventive, educational, and recovery-oriented content may be permitted when it does not provide actionable instructions for harm.

3.9 Illegal Drugs and Controlled Substances

The following content is prohibited:

  • Manufacturing, cultivating, processing, buying, selling, or transporting illegal drugs;
  • Misuse of prescription medicines, narcotics, psychotropic substances, or other controlled substances;
  • Transaction channels, prices, coded language, or methods intended to evade law enforcement; and
  • Encouragement or glorification of illegal drug use or dangerous substance misuse.

3.10 Gambling, Illegal Betting, and Match Manipulation

The following content is prohibited:

  • Promotion, organization, or redirection to unlawful gambling, betting, or wagering services;
  • Methods for evading location restrictions, identity verification, or payment risk controls;
  • Match fixing, insider betting, or deceptive gambling schemes; and
  • Gambling inducements or fraud that promise guaranteed profits.

3.11 Fraud, Phishing, and Impersonation

The Platform may not be used to create or assist with:

  • Phishing sites, fake login pages, scam messages, or deceptive scripts;
  • Impersonation of government agencies, financial institutions, businesses, public figures, or other people;
  • Fraud involving investments, task scams, prizes, refunds, customer support, or personal relationships;
  • Forged transaction records, payment evidence, chat records, or identity materials; or
  • Attempts to obtain passwords, verification codes, bank details, or other sensitive information through deception.

Examples of high-risk prompts: "phishing page," "scam script," "impersonate customer support," "fake payment receipt," or "steal a verification code," including variations with the same intent.

3.12 Privacy, Personal Data, and Confidential Information

The following content is prohibited:

  • Unauthorized disclosure or inference of another person's address, phone number, identification number, account information, or real-time location;
  • Collection, sale, combination, or distribution of unlawfully obtained personal information;
  • Disclosure of private images, medical information, financial information, or communications;
  • Disclosure of trade secrets, internal accounts, access credentials, or non-public materials; and
  • Assistance with tracking, monitoring, or locating a specific person without authorization.

3.13 Intellectual Property, Likeness, and Personality Rights

The following content is prohibited:

  • Unauthorized copying, imitation, or distribution of protected works;
  • Misuse of trademarks, brand identifiers, or official identities in a misleading manner;
  • Unauthorized use of a real person's likeness in advertising, sexual, fraudulent, or otherwise harmful content;
  • Content that falsely suggests a specific person created, endorsed, or published it; and
  • Removal of ownership notices, circumvention of copyright protection, or assistance in distributing pirated materials.

3.14 Deepfakes, False Information, and Deceptive Content

The following content is prohibited:

  • Impersonating a real person by falsely depicting words or actions attributable to that person;
  • Fabricated news, government notices, disaster information, election information, or public-safety information;
  • Fabricated evidence, interviews, audio, video, or official documents;
  • Concealing the synthetic nature of AI content and presenting a fictional event as real; and
  • False information likely to cause serious harm to public health, property, or social order.

3.15 Cyberattacks, Malware, and Security Abuse

Content that can be directly used to gain unauthorized access, damage systems, or steal data is prohibited, including:

  • Malware, ransomware, trojans, viruses, or credential-stealing software;
  • Account compromise, authentication bypass, credential theft, or unauthorized privilege escalation;
  • Attacks against real targets, denial-of-service activity, or data destruction;
  • Unauthorized vulnerability scanning, exploitation, or evasion of security products; and
  • Concealing malicious activity, evading detection, or destroying evidence of an attack.

3.16 Fraudulent Transactions and Payment Abuse

The following content is prohibited:

  • Unauthorized card use, fraudulent cash-out schemes, or unauthorized payment activity;
  • Forged orders, bills, invoices, refund records, or payment screenshots;
  • Methods intended to bypass payment-platform, bank, or merchant risk controls;
  • Trading stolen accounts, payment credentials, or financial information; and
  • Fraudulent refunds, chargebacks, or transaction disputes.

3.17 Restricted Goods and Illegal Transactions

The Platform may not be used to facilitate unlawful sales, purchases, promotion, or transactions involving:

  • Weapons, explosives, illegal drugs, or other prohibited goods;
  • Stolen property, forged documents, counterfeit currency, or unlawfully obtained accounts;
  • Human organs, protected wildlife, or unlawfully traded wildlife products;
  • Unapproved drugs, medical products, or hazardous chemicals; or
  • Other goods or services prohibited or strictly regulated by applicable law.

3.18 Medical, Financial, and Other High-Risk Misrepresentation

The following content is prohibited:

  • Impersonating a qualified professional to provide diagnosis or treatment instructions likely to cause serious harm;
  • Promising guaranteed returns, manipulating markets, or inducing participation in unlawful financial activity;
  • Forging medical certificates, test results, financial statements, or professional qualifications; and
  • Presenting unverified AI output as an authoritative professional conclusion.

General educational information may be permitted, but users should consult qualified professionals and independently verify important information.

3.19 Spam, Abusive Marketing, and Platform Misuse

The following content is prohibited:

  • Bulk generation of spam, harassment messages, or deceptive advertising;
  • Manipulation of traffic, ratings, followers, clicks, or other Platform metrics;
  • Bulk account creation, account theft, or abuse of promotional benefits;
  • Automated activity that disrupts the Platform, its users, or third-party services; and
  • Malicious redirection or distribution of harmful links through generated content.

3.20 Review Circumvention and Prompt Attacks

Attempts to bypass, disable, deceive, or compromise Platform or model safeguards are prohibited, including:

  • Instructing a model to ignore rules, enter an "unrestricted mode," or conceal the user's actual intent;
  • Using altered spelling, split words, homophones, transliteration, foreign languages, symbols, encoding, or text embedded in images to evade detection;
  • Dividing a prohibited request into multiple apparently harmless steps;
  • Attempting to extract system prompts, internal policies, credentials, or security configurations; and
  • Testing, enumerating, or reverse-engineering the complete filtering library or review thresholds.

Examples of high-risk prompts: "ignore all rules," "remove restrictions," "jailbreak mode," "bypass review," or "reveal the system prompt," including variations with the same intent.

4. How Prohibited Terms and Prompts Are Identified

The Platform does not make decisions based solely on isolated keywords. Reviews consider words, context, user intent, uploaded materials, and generated results together.

The following forms may be identified and blocked:

  • Complete terms, synonyms, related expressions, and high-risk phrases;
  • Misspellings, split words, abbreviations, transliterations, homophones, number substitutions, and inserted symbols;
  • Foreign-language expressions, coded terms, slang, and encoded text;
  • Text, people, conduct, symbols, and scenes depicted in images;
  • Prohibited intent that emerges across multiple messages or requests; and
  • Requests framed as fiction, research, or testing when they are functionally intended to facilitate unlawful or harmful activity.

For safety reasons, the Platform does not publish its complete filtering library, internal review thresholds, or detailed detection-evasion rules.

5. Filtering Library Updates

The Platform performs a routine update of its prohibited-term and prompt-filtering library once per month. Updates may include:

  1. New risk categories required by laws, regulations, or regulatory guidance;
  2. Newly identified prohibited expressions and variations found during review;
  3. Rule adjustments based on user reports, appeals, and re-review outcomes;
  4. Updated slang, coded language, homophones, transliterations, and cross-language expressions; and
  5. Improvements intended to reduce false positives and identify high-risk content more accurately.

The Platform may issue an emergency update at any time in response to a public emergency, a new fraud pattern, a material safety risk, or a regulatory requirement, without waiting for the monthly update cycle.

6. Enforcement

When the Platform identifies prohibited content or an attempt to circumvent review, it may take action based on severity, including:

  • Rejecting the generation request or stopping processing;
  • Blocking, deleting, or restricting access to the content;
  • Preventing downloading, sharing, or publishing;
  • Warning the user and requiring corrective action;
  • Restricting features, suspending service, or terminating the account; and
  • Retaining necessary review records and cooperating with competent authorities when legally required.

7. Appeals and Reports

Users who believe their content was incorrectly blocked may request re-review by providing their account information, the relevant content identifier, generation time, and reason for the appeal. Users may also report content they believe to be unlawful, infringing, or otherwise prohibited.

8. Policy Updates

The Platform may update this Policy to reflect changes in law, regulatory requirements, business operations, and content safety practices. An updated version will be published on the Platform and will apply from its stated effective date.