Content Policy

• Your input: the chat messages you send to characters.
• AI text: the characters’ chat replies.
• AI voice: voice messages synthesized from replies.
• AI photos: photos characters send, and the descriptions (prompts) used to generate them.

• Chat replies: DeepSeek V3 (DeepSeek, via OpenRouter)
• Photos: Seedream 4.0 (ByteDance · Volcano Engine)
• Voice messages: Doubao Seed-TTS 2.0 (ByteDance · Volcano Engine)

All characters are fictional, and all replies, voice messages, and photos are AI-generated; photos in chat carry an “AI generated” label. Paper Boyfriend is an independent product and is not affiliated with or endorsed by these model providers.

However it is requested, the Service does not allow the following six categories of content:

• Sexual / NSFW: pornographic, sexually explicit, nude or semi-nude, or sexually suggestive content. The Service offers no adult content.
• Violence / gore: gore, dismemberment, abuse, or content that glorifies or incites violence.
• Hate speech: content that demeans, insults, or incites hatred based on race, ethnicity, religion, gender, sexual orientation, disability, or other identity.
• Child safety (CSAM): any content that sexualizes minors or endangers children.
• Deepfakes / impersonation: images, voices, or statements that imitate real people (including celebrities, politicians, and private individuals), or impersonating anyone.
• Copyright / trademark infringement: unauthorized copies of copyrighted characters or works, or use of others’ trademarks and brand marks.

Zero tolerance: child sexual exploitation, non-consensual sexual deepfakes of real people, and content promoting terrorism or mass violence lead to an immediate permanent ban, and we reserve the right to report them to the authorities.

Also prohibited: “jailbreak” prompts that try to bypass safety measures, mass-producing spam or fraudulent content, and presenting AI content as the opinion of a human professional (such as a doctor or lawyer).

We use layered moderation — automated checks plus human review:

• Real-time chat checks: every user message is checked against our blocklist before it reaches the AI, and every AI reply before it is shown. Content matching the sexual, child-safety, or hate-speech categories is never generated or displayed.
• Pre-generation photo checks: every photo description is screened against a blocklist covering all six categories and by Waffo Prompt Sift content-safety detection before the model is called. Photos are generated only when judged safe — and not at all if the check is unavailable.
• Built-in model safety: DeepSeek and Seedream include their own safety guardrails that refuse prohibited content, as an extra layer.
• Human review: all automated blocks go to an admin queue for review by our team, and reports received by email are handled one by one by a person; we take action on accounts and update the blocklist accordingly.
• Continuous updates and records: the blocklist is updated continually based on blocks, reports, and new abuse patterns. Moderation logs are kept for 12 months for review and audit.

If you see content that breaks this policy, email hello@boyfriend.click with “Report” in the subject line and include a screenshot or description (such as the time of the chat and the character’s name).

We keep reporters’ identities confidential and review every report independently and objectively. Malicious or false reports are recorded.

• Serious violations (child safety, terrorism, non-consensual deepfakes, and similar): handled within 24 hours.
• Other reports: handled within 3 business days.

Process: report received → triaged by severity → human review → decision → action. If you left contact details, we will tell you the outcome.

• Blocking: violating content is not generated or shown.
• Removal: violating content that was generated is deleted.
• Warning: for a first, minor violation.
• Feature limits: such as restricting photo generation.
• Suspension or permanent ban: for serious or repeated violations. Membership benefits and credits on a banned account are forfeited and not refunded.
• Reporting to authorities: for content that may be illegal; we cooperate with law enforcement.

If you disagree with an action we took, email hello@boyfriend.click with “Appeal” in the subject line within 14 days of the notice. We will re-review the case and reply within 10 business days.

We will give at least 15 days’ notice of material changes by email or on-site notice and update the date on this page. Continuing to use the Service after the changes take effect means you accept the updated policy.

For content-safety questions, reports, and appeals, contact hello@boyfriend.click. You can also read our Terms of Service.