How this works

This library was written by software and checked by software. No psychologist signed it off. That is worth knowing before you decide how much weight to give it, so here is the whole of it.

Where the words came from

Every feeling page was written by an AI language model, drawing on the general psychological knowledge in its training rather than a reading list we handed it. Each feeling was put through the same set of questions, which is why any two pages line up closely enough to compare.

Nothing goes live unchecked. Automated checks look for sections that are empty or padded out, language that instructs instead of describes, anything that drifts towards telling you what you have, and links between feelings that point nowhere. Pages that score too low are not published.

How the identify tool picks what it shows you

You describe how you feel, an AI reads it, searches these same pages, and comes back with the two or three that fit best and a short note on why. It chooses from what is already written here rather than inventing something about you, and it stops at naming. It is not assessing you, and it is not speaking for a professional.

What we do not claim

• No licensed clinician wrote or reviewed these pages.
• None of it is peer reviewed.
• The pages cite no individual studies, so they are a vocabulary to start from, not evidence to lean on.

So how much should you trust it

Enough to find a word, loosely enough to put it down again. Where a page and a professional disagree, the professional is right. Where a description does not match how it actually feels, the page is what is wrong, and you are allowed to keep looking. Each feeling page carries the date it was last updated.

How it gets better

We read the descriptions people type into the identify tool, with no name attached to any of them, to find where it misses. That is how we learned it had nothing to offer anyone typing words like numbness or dissociation. What happens to that text, and for how long, is set out on the privacy page.