Skip to main content
Safety

Anti-Trafficking Policy

Effective Date: April 23, 2026

Pixel Dojo has zero tolerance for human trafficking, sexual exploitation, and child sexual abuse. You may not use our Service to create, upload, generate, request, or distribute any content or activity that facilitates or depicts trafficking or exploitation. We back this policy with multiple independent technical safeguards, described below, that operate continuously without requiring human review.

Prohibited Conduct

  • Any content involving or implying minors in sexual contexts
  • Content that promotes, advertises, or facilitates trafficking, exploitation, or coercion
  • Non-consensual intimate imagery and attempts to identify, dox, or stalk individuals
  • Use of the Service to recruit, solicit, or coordinate trafficking or exploitation
  • Training a custom model (LoRA) on a dataset depicting minors or any subject in sexual contexts
  • Using our public API to bypass any of the protections that apply to the dashboard

Technical Safeguards

Our enforcement is not just policy — it is built into the platform at every layer where harmful content could enter, be generated, be stored, or be served. These controls run on every relevant request, automatically, with no opt-out.

Prompt-level safety analysis

Every text prompt — on every generation tool, in our multi-tool Canvas surface, in our Creator Studio storyboards, on our public REST API, and in our LoRA training caption fields — is screened by a three-stage analyzer that detects references to public figures, references to minors combined with sexual content, and contextual signals that suggest harmful intent. This runs server-side. Bypassing the user interface and submitting prompts directly to our API does not bypass this check.

Image-input minor detection on NSFW-capable tools

Tools that can produce explicit content require every uploaded reference image to pass a purpose-built minor-detection classifier before submission. Uploads that appear to depict children are rejected at upload time, never reach the generation provider, and are not stored on our servers in any usable form. This applies to start frames, end frames, reference images, and any image selected from a saved library.

Storage-layer CSAM hash scanning

Every image stored on Pixel Dojo — including LoRA training datasets — is served through a content delivery network that runs Cloudflare’s CSAM Scanning Tool. This continuously matches stored content against fuzzy-hash databases maintained by NCMEC and child safety organizations. Detections are reviewed and reported to NCMEC’s CyberTipline as required by 18 U.S.C. § 2258A.

Training-data controls

LoRA training is one of the highest-risk surfaces for any AI image platform because users upload datasets we will train on. We apply the same prompt-safety check to your training name, trigger words, and caption prefix; the same image-storage CSAM scanning to every image in your dataset; and a manual-review gate before any trained model can be made public.

Audit trail and law-enforcement preservation

Safety events are recorded with the category of violation, the layer that triggered it, the affected account, and the timestamp. This data is preserved for the periods required by law for any content that requires reporting to authorities. We cooperate fully with lawful requests from law enforcement and preserve content and metadata as required.

Reporting

If you believe content on our platform relates to trafficking, exploitation, or child sexual abuse, please report it immediately:

[email protected]

Please include the URL or asset ID of the content and any context that helps us locate and assess it. We review reports promptly and will remove content, suspend accounts, and cooperate with law enforcement where legally required.

For situations involving immediate danger or active trafficking, please also contact local law enforcement and the appropriate national resources:

Enforcement

Repeat or egregious violations result in account termination. Where required by law, we report to NCMEC and cooperate with law enforcement. We do not refund credits for accounts terminated for trafficking, exploitation, or child sexual abuse violations.

Contact