1. Overview
Payoup (operated by Enexfi LTD) implements a comprehensive content moderation framework to ensure the safety, legality, and appropriateness of all user-generated content published on our platform. This policy details our content monitoring, review, and enforcement procedures.
Content moderation decisions are made by our in-house Trust & Safety team. We utilise a third-party AI provider (OpenAI) for automated content classification. No third-party human moderation service is used — all review decisions are made internally.
2. Scope
This policy covers all user-generated content on the platform:
- Payment link titles and descriptions
- Store product listings and descriptions
- Invoice descriptions
- Subscription plan names and details
- User profile information (bio, portfolio)
- Site builder content
- Embedded payment (POS) descriptions
3. Methodology
Our content moderation employs a multi-layered approach combining automated systems with human oversight:
Layer 1: Pre-Publication Automated Filter (Real-time)
All content is automatically scanned at the point of creation and blocked from publication if prohibited content is detected.
Layer 2: Community-Driven Reporting (Post-Publication)
Platform users can report suspicious content through a structured reporting mechanism.
Layer 3: Human Review & Investigation
Trust & Safety team reviews all flagged cases, escalated disputes, and automated alerts.
Layer 4: Periodic Scanning
Existing content is automatically re-scanned every 2 hours against updated policy and prohibited term databases.
4. Automated Systems
Our automated content classification engine includes the following capabilities:
NLP-Based Classification
Content is classified in real-time against 10+ prohibited category taxonomies.
AI Content Analysis (OpenAI)
Third-party AI model (OpenAI Moderation API) analyses text and image content as an additional layer. Categories such as hate speech, violence, adult content, and self-harm are detected.
Image Moderation (AI Vision)
Uploaded product images are scanned via AI-based visual analysis: nudity, weapons, drugs, violence, hate symbols, and gambling content are detected.
Multi-Language Processing
Full Turkish and English support with evasion detection (character substitution, obfuscation, phonetic variants).
Risk-Based Scoring
Each moderation event is assigned a severity level (critical, high, medium, low).
Trust Score Framework
Dynamic trust score system based on user behaviour informs review prioritisation.
5. Human Review
Our in-house Trust & Safety team conducts reviews in the following situations:
- Cases flagged by automated systems that require human judgement
- Content reported through user reporting mechanisms
- Payment links or products receiving multiple reports
- Content from accounts with declining trust scores
- Critical/urgent severity cases
- User appeals against moderation decisions
Review decisions are recorded in an immutable audit log with reviewer identity, timestamp, and rationale.
6. User Reporting Mechanism
All platform users (including customers and payers) can report suspicious content:
In-App Reporting
"Report this link" button is available on every payment link and product page.
Structured Taxonomy
Reports are categorised: scam, prohibited product/service, phishing, misleading information, other.
Abuse Prevention
IP-based rate limiting prevents reporting abuse.
Automatic Escalation
Content receiving 3+ reports is automatically placed under "under review" status.
7. Prohibited Content Categories
Content creation in the following categories is strictly prohibited:
- Illegal Substances — Drugs, controlled substances, and related paraphernalia
- Weapons & Explosives — Firearms, ammunition, explosives
- Counterfeit Documents — Fake IDs, passports, diplomas, and documents
- Adult Content — Pornography, escort services, sexual content
- Gambling — Illegal betting, casino, gambling services
- Fraud & Scams — Ponzi schemes, phishing, money laundering
- Cybercrime — Hacking services, malware, stolen data
- Stolen Goods — Stolen products, accounts, credit card data
- Terrorism — Terror financing, propaganda, radicalisation
- Human Trafficking — Human trafficking, forced labour, organ trafficking
- Prohibited Financial Services — Unlicensed investment, pump & dump, guaranteed profit schemes
8. Escalation Matrix
Moderation cases are managed according to the following priority levels and escalation rules:
| Severity | Trigger | Action |
|---|
| Critical | Terrorism, CSAM, human trafficking, organ trafficking | Content immediately blocked, account permanently suspended, relevant authorities notified where required |
| High | Drugs, weapons, fraud, gambling, adult content, cybercrime, counterfeit documents | Content immediately blocked, user trust score reduced, account placed under review on 3rd violation with all transactions frozen |
| Medium | 3+ user reports on same content | Content deactivated and cannot accept payments, user notified, trust score reduced by 10, awaiting admin review |
| Low | Single user report | Content remains active, added to review queue, evaluated by Trust & Safety team with normal priority |
9. Enforcement Actions
Graduated enforcement model applied based on moderation decisions:
Content Blocking
Prohibited content is automatically blocked pre-publication.
Content Removal
Published content found in violation is removed.
Trust Score Reduction
User trust score is automatically reduced with each violation.
Account Under Review
Account is placed under review after 3+ repeated violations.
Account Suspension
Account is suspended for serious or repeated violations.
Account Termination
Accounts with zero trust score are permanently removed from the platform.
Legal Referral
Critical-level violations (terrorism, CSAM) are reported to relevant authorities.
10. SLA & Response Times
| Trigger | Action | Target SLA |
|---|
| Automated block (pre-publication) | Content rejected | Real-time |
| 3+ community reports | Automatic under-review status | Immediate |
| Human review (standard) | T&S team investigation | 24–48 hours |
| Critical/Urgent cases | Urgent review | < 4 hours |
| Appeal review | Senior review | 5 business days |
11. Appeals Process
A formal appeals process is available for users who wish to contest moderation decisions:
- Appeals must be submitted within 14 days of the moderation decision.
- User must provide the appeal reason and supporting information.
- Appeals are evaluated by a senior reviewer different from the original decision maker.
- Appeal outcomes are communicated within 5 business days.
- Approved appeals result in content restoration and trust score reinstatement.
12. Audit & Transparency
All moderation activities are subject to a comprehensive audit framework:
Immutable Audit Trail
Every moderation decision (block, approval, rejection, escalation) is recorded with timestamp, reviewer identity, and rationale.
Daily Metrics
Total scans, auto-blocks, human reviews, approvals, rejections, and average resolution time are tracked daily.
SLA Monitoring
Resolution time for each case is measured and compared against target SLAs.
Periodic Review
Moderation policies and prohibited term database are updated regularly.