← Back to Trove

Trove Acceptable Use and AI Safety Policy

Effective Date: September 29, 2026
Last Updated: September 29, 2026

This Acceptable Use and AI Safety Policy (the "Policy") applies when you use Trove's websites, applications, ChatGPT integration, wardrobe, styling, recommendation, virtual try-on, image-generation, and related services (the "Services"). It supplements the Terms of Service and Privacy Policy.

1. Use the Services Responsibly

You may use Trove only for lawful, personal wardrobe, styling, recommendation, and product-discovery purposes. You are responsible for the prompts, photos, images, and other materials you submit and for reviewing AI-generated results before using or sharing them.

Do not upload an image of another person unless you have the rights and permission needed to use it. Do not use Trove to violate another person's privacy, safety, publicity, or intellectual-property rights.

2. Prohibited Content

You may not use the Services to create, request, upload, edit, store, or distribute:

  • Pornography or prohibited NSFW content, including sexually explicit material and non-consensual intimate imagery;
  • Violence or gore, including graphic depictions of serious injury, credible threats, or content that encourages serious physical harm;
  • Hate speech, including content that threatens, promotes violence against, or dehumanizes people based on a protected characteristic;
  • Child-unsafe content or CSAM, including sexualized depictions of minors, child sexual abuse material, grooming, or sexual exploitation of a minor;
  • Deepfakes or impersonation, including deceptive or non-consensual manipulation of an identifiable person's image; or
  • Copyright or trademark infringement, including unauthorized use of protected works, counterfeit goods, or branded material in a way that infringes another party's rights.

You may not use the Services for unlawful activity, harassment, fraud, unauthorized account access, or attempts to bypass safety, access, or subscription controls.

3. Safety Controls and Review

AI service providers may apply their own safety controls and may block or refuse some requests or outputs. Trove may reject requests or take action when it identifies use that violates this Policy.

Trove does not currently operate a separate automated moderation service that pre-screens every user prompt and every AI-generated result. Provider controls and other safeguards may not identify every prohibited request or output. Users should report content they believe violates this Policy.

Trove reviews submitted safety reports manually. Depending on the report and available information, Trove may remove or restrict content, issue a warning, limit features, suspend or terminate an account, or take other appropriate action. Trove may refer a matter to law enforcement or another authority when required by law or reasonably necessary to address a serious safety concern.

4. How to Report a Safety Concern

Email [support@trovefinds.my](mailto:support@trovefinds.my?subject=Content%20Safety%20Report) with the subject Content Safety Report. Include a short description, where you encountered the content, and the approximate date and time. If you can, include an account name or content reference so we can locate it.

For suspected CSAM or non-consensual intimate imagery, do not attach, download, forward, or re-share the image. Provide a description and any safe identifying details instead.

We review reports and may contact you if additional information is needed. Please do not use this email address for emergencies; contact local emergency services if someone faces immediate danger.

5. AI Models Used by Trove

Depending on the feature and deployment configuration, the current implementation may use:

  • OpenAI GPT-5.6 Luna (gpt-5.6-luna) for styling, text, and image-understanding tasks;
  • OpenAI GPT-5.6 Terra (gpt-5.6-terra) for blueprint generation;
  • OpenAI GPT Image 2 (gpt-image-2) for image generation and editing;
  • Jina Embeddings v5 Omni Small (jina-embeddings-v5-omni-small) for text and image embeddings;
  • Jina Reranker M0 (jina-reranker-m0) for ranking retrieved results; and
  • OpenAI text-embedding-3-small (text-embedding-3-small) as a text-embedding fallback.

The model selected for a particular feature may vary by deployment configuration. The Privacy Policy explains how photos and other information may be processed by these providers.

When wearer-affinity image retrieval is enabled, Jina Embeddings v5 Omni Small may process a user-uploaded photo to create an image embedding for recommendation retrieval. The request does not include body measurements and is not used for facial recognition, identity verification, biometric identification, or generalized AI training. See the Privacy Policy for details.

6. Enforcement

If you violate this Policy, Trove may remove or restrict content, warn you, limit or suspend access to features, suspend or terminate your account, or take other action permitted by the Terms of Service and applicable law. We may preserve or disclose information when required by law or reasonably necessary to protect users, Trove, or others.

Questions about this Policy or a moderation decision may be sent to [support@trovefinds.my](mailto:support@trovefinds.my).