How it's made.

The quotes. Every line is checked word for word against its original, public-domain text, and each page links to the text it was checked against. Where the famous version is a misquote, the page says so. The plain-words note beside it is our reading of the text, not advice.

The stories. Every story here starts with something real — a memory, a person, a feeling I've carried. Names, details, and settings are changed to protect the people I love. Nothing here is trying to teach you something. It's just a moment, held a little longer than usual. The read-aloud voice is an AI voice, from smallest.ai.

Talk. When you talk here, you are talking to an AI. Talk runs on Claude, a model made by Anthropic. It says so itself: it calls itself Yohaanei AI and never pretends to be a person. It listens, names what's underneath, and gives advice only if you ask. It is not a therapist or a doctor.

What you write goes to Anthropic's API, Claude replies, and your words are gone. We never store them, and Anthropic does not train on them by default (the details are in the privacy policy). If you turn on memory in your account, we keep a sentence or two about the theme of a conversation, never your words. It is off unless you choose it.

Before any message reaches Claude, our own code checks it for signs of crisis, distress or harm to a child. If it finds one, you see free helplines in India and where to find one elsewhere, even if Claude is down.

Behind the scenes. We use AI tools for the work around the writing — first drafts, searching old texts, checking our own pages. A quote is never made up by one: it is on the page only once it has been found in the original.