Rosely Private Limited ("Rosely", "we", "us", or "our") takes content safety, legal compliance, and platform integrity seriously. To maintain a secure and responsible environment on Rosely.ai, we have implemented a dual-layer content moderation system comprising pre-screening (before content is generated or displayed) and post-screening (after content has been generated or made available).
This policy describes our commitment to reviewing content both proactively and reactively to ensure it aligns with our community standards, Prohibited Content Policy, and applicable law.
1. Content Subject to Screening
The following types of content are subject to both pre-screening and post-screening review:
- All AI-generated text, images, video, and other outputs
- User-submitted text prompts and chat inputs
- Account metadata including usernames, profile information, and any user-configured companion attributes
2. Pre-Screening
Before any content is generated or displayed, the following measures are applied:
Automated Input Filtering: All user prompts and inputs are scanned in real time by automated safety classifiers trained to detect prohibited content, high-risk language patterns, and policy violations. Inputs that match prohibited categories are blocked before any AI output is generated. Pattern-Based Blocking: Our systems maintain and regularly update a list of blocked terms, patterns, and behavioral signals associated with prohibited content. These filters are applied to all inputs regardless of how they are framed. Contextual Analysis: Beyond keyword matching, our systems assess the contextual intent of inputs. Attempts to circumvent filters through indirect phrasing, coded language, or gradual escalation are specifically targeted. This pre-screening layer is designed to prevent prohibited content from being generated in the first place.
3. Post-Screening
After content is generated, additional review measures are applied:
- Automated Output Scanning: All AI-generated outputs are reviewed by our content classifiers before delivery to the user. Outputs that trigger our prohibited content detectors are withheld and flagged for review.
- Continuous Monitoring: Our automated systems continuously monitor platform activity for emerging patterns or violations that may not have been caught at the input stage.
- Human Moderation: Our moderation team conducts regular audits of flagged content, trending patterns, and user-submitted reports. Human moderators have authority to remove content, adjust model behavior, and escalate accounts for enforcement.
- User Reporting: Users can report any content for further evaluation at any time by emailing support@rosely.ai.
- We reserve the right to remove, restrict, or take enforcement action on any content at any time, including content that has already been delivered to a user, if it is found to violate our policies or applicable law.
4. Enforcement
Content found to be in violation during either screening phase may result in:
- Immediate removal or blocking of the output
- Suspension or permanent termination of the user’s account
- Escalation to law enforcement or relevant authorities where legally required
- Preservation of evidence for legal proceedings
5. Continuous Improvement
Our moderation systems are regularly updated to reflect evolving standards, emerging abuse patterns, new legal requirements, and improvements in detection technology. We treat content safety as an ongoing operational commitment, not a static checklist.
6. Contact Us
Questions or reports:
- Email: support@rosely.ai
- Website: https://rosely.ai