Editorial standards
How the Unrestricted blog sources its claims, checks facts, writes about its own product, labels updates and corrects mistakes.
The Unrestricted blog explains how AI chat products work: why assistants refuse, what their policies allow, what they keep, and how uncensored models are made. These are the rules every post follows.
Sources
Every claim about another company's product or policy, a law, or a statistic links to a primary source: the company's own help pages and policies, the paper itself, the model card from its publisher, or the statute or agency page. Sources are listed at the end of each post with the month we checked them. We don't average figures from unnamed websites, and we don't ask an AI model for a fact or a figure.
No invented data
We don't publish made-up statistics, surveys, studies, quotes, reviews or user counts. If a post reports a test we ran, it names the model, version, plan and date and links the saved transcripts. No transcript, no claim.
How posts are made
We use AI tools to help research and draft posts. Before a post is published, each factual claim is checked against the source it cites, and anything we can't support is removed or softened. Posts are bylined "Unrestricted team" unless a named person reviewed them.
Writing about our own product
What a post says about Unrestricted is checked against how the product actually works today. When the product changes, we update the posts that describe it.
Updates and corrections
A post shows "Updated" with the month when its content changes. We re-check posts about other companies' policies regularly, because those policies change often. When we get something wrong, we fix it and say so in a note at the top of the post. In November 2026 we corrected ten earlier posts and removed seven that relied on figures and product claims we could not support.
Spotted an error? Email hello@unrestricted.chat.
Not advice
Posts are general information. They are not legal, medical or financial advice.