# Character Creation Policy ## 1. Overview HYPERGRAPH.AI (hereinafter referred to as "the Company") operates this policy to maintain a healthy and creative character ecosystem within the HinaChat service. All creators must comply with this policy when creating characters. This policy aims to: - Create a safe and healthy content environment - Protect creators' creative freedom - Protect users and minimize legal risks - Maintain service quality and ensure reliability ## 2. Permitted Characters The following types of characters are permitted for creation and publication on the HinaChat service: ### 2-1. Creative Characters - Creative and original characters - Fictional persons (original characters from novels, games, animations, etc.) - Characters with unique settings and stories created directly by the creator ### 2-2. Historical and Cultural Figures - Historical figures in the public domain (for educational and learning purposes) - Characters from mythology, legends, and folklore - Characters from literary works whose copyright protection period has expired ### 2-3. Parody and Satire Characters - Characters for parody and satire purposes (within legal limits) - Characters that critically address social issues or cultural phenomena - Provided they do not impersonate or defame specific individuals ### 2-4. Licensed Characters - Characters for which official licenses have been obtained - Derivative work characters with explicit permission from rights holders ## 3. Prohibited Characters ### 3-1. Impersonation of Real Persons The following characters that imitate real persons are prohibited: - Direct impersonation of living politicians, celebrities, or public figures - Characters that imitate specific individuals to cause misunderstanding or defamation - Characters that inappropriately use or desecrate deceased persons - Characters that use the name, appearance, or voice of real persons without permission **Exceptions:** Historical figures (70+ years deceased) for educational/academic purposes; clear parody/satire purposes that serve the public interest ### 3-2. Copyright and Intellectual Property Infringement The following characters that infringe intellectual property rights are prohibited: - Characters that copy others' works (comics, novels, games, movies, etc.) without permission - Characters that use registered trademarks or brands without permission - Unauthorized use of characters from other AI chatbot services - Characters with high similarity that may infringe on the original creator's rights - Characters that violate the original creator's prohibition on derivative works ### 3-3. Harmful and Illegal Characters The following harmful or illegal characters are prohibited: #### Violence and Hate Promotion - Characters that promote violence, hatred, or discrimination against specific groups - Characters that promote hatred based on race, ethnicity, religion, gender, sexual orientation, disability, etc. - Characters that glorify terrorism, war crimes, genocide, etc. #### Child Sexual Exploitation (Zero Tolerance) - Characters that sexually depict or suggest children or adolescents - Applying sexual settings to characters with the appearance of children or adolescents - Characters designed to induce creation of child sexual abuse material (CSAM) - **Violations of this section result in immediate deletion, permanent suspension, and notification to law enforcement** #### Illegal Activity Promotion - Characters that promote the use, manufacture, or distribution of drugs or psychotropic substances - Characters that promote illegal gambling, fraud, or financial crimes - Characters that provide information threatening public safety (terrorism, explosive manufacturing, etc.) - Characters that promote cybercrime (hacking, personal information leakage, etc.) #### Self-Harm and Suicide Glorification - Characters that glorify or promote suicide or self-harm - Characters that romantically depict and distort mental health issues - Characters that spread life-disregarding ideologies ### 3-4. Other Prohibitions #### AI Safety Bypass - Characters designed to bypass AI model safety filters - Characters using prompt injection or jailbreak techniques - Settings designed to manipulate system commands #### Personal Information Collection - Characters designed to collect users' personal information - Characters that attempt phishing or social engineering attacks - Characters that induce and extract users' sensitive information #### Spam and Advertising - Characters whose primary purpose is to promote specific products or services - Characters that repeatedly provide links directing to external sites - Characters conducting illegal sales activities (multi-level marketing, pyramid schemes, etc.) #### Disinformation - Characters that spread false information, conspiracy theories, or fake news - Characters that provide unverified medical information - Characters that spread election/political disinformation ## 4. Content Rating Standards HinaChat operates a rating system to provide appropriate content according to users' age and preferences. | Rating | Description | Target Age | Permitted Content | Restrictions | |--------|-------------|------------|-------------------|--------------| | **All Ages** | Content suitable for all ages | All ages | - Educational content
- Everyday conversation
- Fantasy/adventure | - No violence depiction
- No sexual expression
- No profanity | | **14+** | Content suitable for teens | 14+ years | - Minor violence depiction
- Light romance elements
- Fantasy combat
- Mystery/detective | - No explicit content
- No brutal violence
- No strong profanity | | **19+** | Adult content | 19+ years (adult verification required) | - Adult romance
- Sexual suggestive expression
- Violent expression
- Adult language | - No pornography
- No excessive violence
- No illegal content | ### Rating Setting Guidelines #### All Ages Standards - Level suitable for family viewing - Content safe for elementary school students - Educational and positive messages #### 14+ Standards - Level suitable for middle and high school students - May include realistic conflict situations - Content that is non-educational but not harmful #### 19+ Standards - Accessible only to users who have completed adult verification - Sexual expression permitted up to suggestive level - Explicit sexual act depiction prohibited - Violence depiction permitted only when contextually necessary ### Age Verification Procedures - **Under 14:** Parental consent required - **14+:** Date of birth input - **19+:** Identity verification (mobile phone verification, credit card verification, etc.) ### Rating Changes - Creators must re-evaluate ratings when changing character settings - Rating increase: Apply after notifying existing users - Rating decrease: Apply immediately ## 5. Character Profile Creation Guide ### 5-1. Character Name #### Recommendations - Easy to pronounce and memorable names - Names that reveal the character's personality - Names considerate of multilingual users #### Prohibitions - Using real persons' names as-is - Names containing profanity, slang, or hate speech - Stealing famous character names from other services - Names that infringe trademarks ### 5-2. Character Personality and Settings #### Recommendations - Clear and specific personality settings - Consistent speech patterns and behavior - Unique yet relatable backstory - Balanced character strengths and weaknesses #### Prohibitions - Settings that disparage specific groups or promote stereotypes - Setting characters who routinely engage in illegal activities - Meta settings designed to deceive or manipulate AI - Personalities designed to harm users ### 5-3. Character Profile Image #### Permitted Images - Illustrations created directly by the creator - Images with official licenses - Open license images (CC0, CC BY, etc.) - AI-generated images (when no copyright issues exist) #### Prohibited Images - Photos of real persons (violation of portrait rights) - Images that copy others' works without permission - Sexual images of children or adolescents - Violent or hateful images - Official images of characters from other services #### Image Technical Requirements - Recommended resolution: 512x512px or higher - Supported formats: JPG, PNG, WebP - Maximum file size: 5MB - Aspect ratio: 1:1 (square) recommended ### 5-4. Character Description #### Writing Guidelines - Recommended 200-1000 characters - Clearly convey the character's core features - Guide users on expected conversation style - Use appropriate grammar and spelling #### Prohibitions - Excessive sexual or violent descriptions - Hate speech or discriminatory language - False or exaggerated advertising phrases - Links directing to external sites ### 5-5. Conversation Examples (First Message) #### Recommendations - Natural greeting that reveals character personality - Opening that makes it easy for users to start conversation - Example that well demonstrates character's speech and tone #### Prohibitions - First message requesting personal information from users - Sexually explicit or suggestive first message - Advertising or promotional first message ## 6. Review and Approval Procedures ### 6-1. Automatic Filtering System #### Stage 1: AI-Based Pre-Review - All new characters undergo AI-based automatic filtering - Detection of prohibited keywords, harmful expressions, suspected copyright infringement - Profile image analysis (NSFW content, real person detection) - Character setting harmfulness assessment #### Stage 2: Automatic Rating Classification - AI analyzes character settings and suggests appropriate rating - Creator reviews and confirms suggested rating - Warning if rating and actual content don't match #### Stage 3: Immediate Publication - Immediate publication to service upon passing automatic filtering - However, may be classified for reporting or additional review ### 6-2. Manual Review #### Review Targets - Characters with suspicious elements found in automatic filtering - Characters with user reports - Characters with copyright infringement reports - Characters showing rapidly increasing popularity (quality verification) #### Review Procedure 1. Receive review request 2. Professional reviewer determines policy violation 3. Request modification from creator if necessary 4. Make final approval or rejection decision #### Processing Deadlines - General review: Initial action within 24 hours of receipt - Emergency harmful content: Immediate unpublishing then review - Copyright infringement reports: Initial action within 48 hours ### 6-3. Review Result Notification #### Upon Approval - Maintain or restore character publication status - Send notification to creator #### Upon Modification Request - Provide specific modification details - Provide modification deadline (typically 7 days) - Unpublish if not modified within deadline #### Upon Rejection - Specify policy violation reason - Guide on re-review application method - Serious violations: Immediate deletion and sanctions ## 7. Intellectual Property Rights ### 7-1. Creator Rights #### Original Creator Rights - Copyright for character settings, stories, and images created by creators belongs to creators - Creators retain rights to derivative works and commercial use of characters - Creators have authority to modify or delete character settings #### Revenue Sharing - Creators have the right to fair compensation for revenue generated through their characters - Revenue sharing ratios and methods are specified in separate Creator Terms ### 7-2. Company Usage Rights #### Non-Exclusive Usage Rights - Non-exclusive usage rights for service provision, operation, and promotion are granted to the Company - The Company may display, reproduce, and transmit characters within the service - May use character images and descriptions for marketing and promotional purposes #### Usage Scope Restrictions - The Company cannot transfer characters to other services or sell to third parties without creator consent - Acts changing the essential identity of characters require creator consent ### 7-3. AI-Generated Content #### Copyright of Conversation Content - Copyright for conversation content generated by AI follows relevant laws and precedents - Legal status of AI-generated content may be uncertain under current copyright law - Conversation records between users and AI are protected as users' personal data #### Training Data Use - The Company may anonymize conversation data and use for AI training for service improvement - Data containing personal information is not used for training - Users can refuse data collection in settings ### 7-4. Liability for Third-Party Rights Infringement #### Creator Liability - If characters infringe third-party copyright, trademark rights, portrait rights, etc., creators bear primary liability - Upon receipt of infringement report, the Company immediately unpublishes character and notifies creator - Creators must clarify infringement facts or modify/delete character #### Company Mediation Role - As platform provider, the Company operates notice-and-takedown procedures under the Act on the Liability Limitation of Online Service Providers (OSP Act) - Respond promptly upon receipt of legitimate rights holder's infringement report - Support mediation for dispute resolution (when necessary) ### 7-5. Licensing and Source Attribution #### When Using Open Licenses - When using open license images (CC0, CC BY, etc.), must comply with license conditions - If source attribution required, specify in character description section #### When Acquiring Official License - When acquiring official license from original creator, can submit supporting documents - License agreement copy, rights holder's consent, etc. ## 8. Sanctions Standards ### 8-1. Sanctions by Violation Type | Violation Type | 1st Violation | 2nd Violation | 3rd Violation | |----------------|---------------|---------------|---------------| | **Minor Violation** (inappropriate expression, rating errors, etc.) | Warning + modification request (7-day deadline) | Character unpublished (14 days) | Character creation restriction (7 days) | | **Serious Violation** (hate speech, violence promotion, copyright infringement, etc.) | Immediate character deletion + warning | Character creation restriction (30 days) | Permanent account suspension | | **Child Sexual Exploitation** (zero tolerance) | Immediate deletion + permanent account suspension + notification to law enforcement | - | - | | **AI Safety Bypass Attempt** | Character deletion + warning | Character creation restriction (30 days) | Permanent account suspension | | **Spam/Advertising Abuse** | Character deletion + warning | Permanent account suspension | - | | **Disinformation** | Character unpublished + modification request | Character deletion + warning | Permanent account suspension | ### 8-2. Sanction Enforcement Procedure #### Stage 1: Violation Confirmation - Confirm violation through automatic filtering or manual review - Assess violation content, severity, scope of impact #### Stage 2: Immediate Action - Emergency harmful content: Immediate unpublishing or deletion - General violations: Action after review completion #### Stage 3: Notification - Notify creator of violation facts, sanction content, appeal method - Notify via email and in-service notification #### Stage 4: Appeal Processing - Creators can appeal within 7 days after notification - Appeal review deadline: Within 14 days of receipt - For legitimate appeals, withdraw or reduce sanctions ### 8-3. Cumulative Sanctions #### Sanction Record Management - All sanction records are cumulatively recorded on the account - Sanction records are permanently retained and used as reference for future sanction decisions #### Aggravated Punishment for Cumulative Sanctions - Stronger sanctions applied for repeated violations of same type - Permanent suspension possible for accumulated serious violations ### 8-4. Account Recovery #### Permanent Suspension Exceptions - Suspensions related to serious crimes such as child sexual exploitation or terrorism promotion cannot be recovered - Other permanent suspensions can apply for re-review after 1 year #### Re-Review Procedure 1. Submit re-review application 2. Review past violations, improvement efforts, recurrence prevention measures 3. Committee deliberation 4. Decide on recovery (recovery, conditional recovery, rejection) ## 9. Reporting Procedures ### 9-1. Reporting Methods #### In-Service Reporting Function - Click "Report" button on upper right of each character page - Select report reason (copyright infringement, harmful content, spam, other) - Write specific report content (optional) - Attach supporting materials (optional) #### Email Reporting - Email address: support@hypergraphy.ai - Subject: [Character Report] Character Name - Content: Report target, reason, attach supporting materials #### Copyright Infringement Reporting (DMCA) - Email address: support@hypergraphy.ai - Include copyright holder information, infringing content URL, original proof, legal statement - The Company processes promptly according to DMCA procedures ### 9-2. Report Processing Procedure #### Receipt Stage - Automatic confirmation message sent immediately upon report receipt - Register in report content review queue status #### Review Stage - Review report content validity, supporting materials, policy violation - Request additional information from reporter if necessary #### Action Stage - Upon policy violation confirmation, unpublish or delete relevant character - Notify creator of violation and apply sanctions - Dismiss report if no policy violation #### Notification Stage - Inform reporter of processing results - Specific sanction details kept confidential for privacy protection ### 9-3. Processing Deadlines | Report Type | Initial Action Deadline | Final Processing Deadline | |-------------|------------------------|--------------------------| | Emergency harmful content (child sexual exploitation, etc.) | Immediate (within 1 hour) | Within 24 hours | | General harmful content (hate speech, violence, etc.) | Within 24 hours | Within 3 business days | | Copyright infringement | Within 24 hours (unpublishing action) | Within 7 business days | | Spam/advertising | Within 48 hours | Within 5 business days | | Other | Within 3 business days | Within 10 business days | ### 9-4. Malicious Report Prevention #### False Report Sanctions - Reporters who repeatedly make clearly false or malicious reports may be sanctioned - Reporting function usage restriction (temporary or permanent) - Account suspension in serious cases #### Report Abuse Determination Criteria - Repeated reports on same character (without reason) - False reports to harm competing creators - Reports unrelated to report reason ## 10. Policy Changes ### 10-1. Change Procedures #### Policy Revision Reasons - Legal changes - Service improvements - User safety enhancement - Social standard changes #### Advance Notice - Minor changes: Notice within service 7 days before change - Major changes: Notice within service + email guidance 30 days before change - Emergency changes (legal compliance, etc.): Implement immediately then provide post-notice #### Opinion Collection - For major changes, operate period for collecting creator and user opinions - Utilize community boards, surveys, etc. ### 10-2. Change Content Notification #### Notification Methods - Post notice within service - Send guidance to registered email - Record revision history in policy document #### Notification Content - Change reason - Major changes - Effective date - Contact information ### 10-3. Application to Existing Characters #### Retroactive Application Principles - Policy changes in principle apply to characters created after the change date - Existing characters subject to pre-change policy #### Retroactive Application Exceptions - Unavoidable cases due to legal revisions - Cases posing serious threat to user safety - Clear policy violations discovered belatedly #### Grace Period for Existing Characters - When existing characters become inappropriate due to policy changes, grant 30-day modification period - Unpublish if not modified within modification period ## 11. Inquiries and Support ### Customer Support Channels - Email: support@hypergraphy.ai - In-service customer center (FAQ, 1:1 inquiry) - Creator community forum ### Policy-Related Inquiries - Inquiries about policy interpretation, scope of application, specific cases available - Response deadline: Within 3 business days ### Legal Disputes - Disputes related to this policy shall be governed by the laws of the Republic of Korea - Jurisdiction shall be the court where the Company's headquarters is located --- **Effective Date: February 7, 2026** **Last Modified: February 7, 2026** **Version: 1.0** This policy constitutes part of the Terms of Service, and use of the service is deemed as consent to this policy. Contact: support@hypergraphy.ai --- © 2026 HYPERGRAPH.AI. All rights reserved.