Content Moderation
Every piece of posted content passes through a two-layer review — automated systems (AI) and human review — to confirm its integrity. Cases where judgment is divided are always reviewed by a person.
A two-layer system
Automated screening at posting
Images, video, text, and audio are assessed by machine-learning models. Content highly likely to be in violation is blocked immediately, while content whose assessment is inconclusive moves to review by our specialist team. The models are continuously improved in-house. We have also systematized a correction flow for confirmed false positives, operating so that no particular creator remains persistently disadvantaged.
Review by a specialist team
A dedicated moderation team reviews content whose assessment was held by the automated systems, as well as cases reported by users. Decisions follow written guidelines, not individual intuition. Every decision is recorded, and consistency is checked monthly. We also operate safeguards for reviewers' mental health, including rotation and access to counseling.
Decision flow
- 01
Receiving a post
The moment a creator submits a post, the automated systems make an initial assessment.
- 02
AI assessment
Clear violations are blocked immediately, clearly unproblematic content is published, and inconclusive cases move to review by the specialist team.
- 03
Human review
Specialist reviewers evaluate held and reported cases, deciding — in line with the guidelines — whether to publish, request changes, or unpublish.
- 04
Notifying the creator
We inform the creator of the outcome and, where appropriate, the reasons behind it.
- 05
Appeals
Creators who disagree with a decision can file an appeal. A different reviewer re-evaluates the case.
- 06
Records and review
Every decision is recorded, then aggregated and analyzed monthly. Patterns of bias or false positives feed back into improvements to the guidelines.
Main prohibited categories
Content falling under the following categories is never permitted, under any circumstances. Detailed definitions and criteria are set out in our guidelines for creators.
Content involving minors
Filming or publication without consent
Violent or gruesome expression
Discrimination and hate speech
Expression relating to illegal drugs
Fraud or use for solicitation and dating purposes
Copyright or portrait-right infringement
Any other post that may violate the law
Reporting
Reports about content are accepted at all times via the in-service reporting form or our contact desk. For urgent cases, we aim to make an initial response within 24 hours.
