Pre-Generation Multilingual Safety Guard
DigitalXLR deploys a deterministic pre-generation safety filter that inspects text prompts and reference image inputs prior to queue submission. The filter blocks explicit sexual content, nudity, or unauthorized undressing instructions across major languages (including English, Malay, Chinese, Japanese, Korean, Arabic, Russian, Spanish, French, German, Vietnamese, Thai, Hindi, and Tagalog).
When a Prompt Is Blocked
If a prompt triggers safety parameters, the system halts generation and displays a localized safety notice:
"Sorry, I can't generate or edit images involving nudity or explicit sexual content."How to Resolve Safety Rejections
- Simplify the Prompt: Remove exaggerated, violent, or explicit descriptive terms.
- Focus on Neutral Cinematic Language: Describe visual movement using cinematic phrasing (e.g. "The protagonist holds a weapon in hand" instead of violent terminology).
- Use Agent Rewriting: Ask Agent Mode to rewrite your request into a clear, direct prompt that preserves your creative intent while complying with content safety policies.