Proposed effective date: July 30, 2026
Last updated: July 30, 2026
TruePair uses content moderation to support an adults-only, relationship-focused service and enforce the Terms of Use and Community Guidelines.
1. Content covered
Moderation may apply to profile text, prompts, captions, photos, videos, voice notes, messages, attachments, reports, support submissions, account signals, and other content or conduct made available through the Service.
Private communications are not treated as public posts, but they may be processed to deliver the feature, apply safety controls, investigate a report, prevent abuse, or comply with law as described in the Privacy Policy.
2. Automated safeguards and human review
TruePair may use deterministic rules, integrity signals, and third-party automated moderation services to identify likely prohibited content or elevated risk. The release candidate identifies OpenAI’s Moderations API as an automated moderation processor for supported text and media checks under TruePair’s configured data controls.
Automated systems can make mistakes. Human reviewers are required for the moderation states, appeals, and safety workflows defined by the active policy and product behavior. Authorized reviewers receive only the information reasonably necessary for their task.
3. Pending, approved, rejected, and review states
New or edited text and media may be held from publication while automated checks or human review are pending. Content must not become visible merely because a provider is unavailable, times out, returns a malformed response, or cannot confidently classify it.
Depending on the applicable rule and confidence, content may be:
- allowed;
- held as pending while checks continue;
- sent for human review;
- rejected or removed;
- restricted to reduce exposure;
- associated with a warning, feature limit, or account action; or
- escalated into a safety investigation.
The user-facing state should not claim that content is live or approved when it is still pending.
4. Prohibited-content priorities
High-priority categories include apparent child sexual exploitation, credible threats, non-consensual intimate content, coercion, trafficking, severe harassment, hate, violent harm, scams, impersonation, sexual solicitation, and attempts to evade a safety control. Other violations include spam, unauthorized automation, intellectual-property infringement, privacy violations, and content incompatible with TruePair’s intended adult relationship service.
Context, severity, repetition, intent, likely harm, and legal obligations may affect the response.
5. Provider failures and recovery
If an automated provider fails, unreviewed content should remain unavailable until a successful retry or authorized human decision. Retry and recovery operations should be bounded, auditable, and protected against duplicate publication or duplicate enforcement.
TruePair may temporarily limit uploads, edits, or messages when moderation cannot operate safely. Availability does not override the obligation to keep unapproved content from publication.
6. Enforcement
Enforcement may include content removal, reduced visibility, warnings, feature restrictions, account suspension, account removal, preservation of narrowly relevant evidence, or referral to competent authorities when appropriate. TruePair may act without advance notice when necessary to address urgent risk, preserve service integrity, or comply with law.
7. Requesting review or appeal
When the app offers retry, review, or appeal, follow the displayed instructions. You may also contact support@truepair.com with the public content, case, or decision reference and enough information to locate the affected account or content.
An appeal should explain why the decision may be incorrect and provide relevant context. Do not send passwords, one-time codes, payment credentials, full identity documents, or unrelated private information.
An appeal does not guarantee restoration. TruePair may decline duplicate, abusive, fraudulent, untimely, or already-final requests. A reviewer should not decide an appeal merely by repeating the original automated result when designated human review is required.
8. Notice and privacy
TruePair may provide a general reason category and available next steps. We may withhold details that would expose another person’s report, private evidence, security controls, confidential detection methods, or information restricted by law.
Moderation inputs, decisions, appeals, access, retention, and deletion are governed by the Privacy Policy and applicable holds. TruePair does not use moderation data to make employment, credit, housing, education, insurance, or similarly significant decisions.
9. Contact
Moderation review and appeals: support@truepair.com
Safety concerns: safety@truepair.com
Privacy questions: privacy@truepair.com
TruePair is operated by CYBEXO INC., 10 Times Sq, Fl 5 Ste 6143, New York, NY 10018.