Frequently Asked Questions

It is a comprehensive technical check of a chatbot's behavior. We test how it handles unrestricted prompts, whether it maintains character consistency, and if the platform imposes hidden filters not mentioned in their marketing.

A truly uncensored model (or "unfiltered") is an LLM that does not trigger a refusal response ("I can't assist with that") regardless of the scenario's complexity or nature.

The primary risks are data leaks and "shadow-censorship," where a platform may suddenly tighten filters, making a paid subscription useless. We audit these risks so you can choose platforms with stable policies.

Yes. We document current prompt engineering techniques (jailbreaks) for restricted platforms like Character.ai or Poe, allowing users to understand the technical limitations of proprietary filters.

The AI industry moves fast. We re-test major platforms every 30 days to ensure our "Censorship Depth" metrics and technical stats are up to date.