This app is built on AI, and we think you should understand plainly how it's used — what it does, what it doesn't, and what happens to what you type. No jargon, no fine print.
What's actually happening
When you have a conversation in the app, your messages are sent to a large language model — Claude, made by Anthropic — which generates the responses you see. It's the same kind of technology behind other AI assistants you may have used.
What makes this app different is what the model is working from. We've given it direct access to the content of Melis's books, so its answers are grounded in what she actually wrote rather than in a general model's best guess. That's why the app can point you to real passages and page references — it's answering from the source, not improvising.
What we do
- Send your messages to Anthropic so the model can generate a response.
- Ground those responses in Melis's books, and show you the sources.
- In the Collaborator tier, and only if you've turned it on, remember useful context about your work to make the app more helpful.
What we don't do
- We don't use your conversations to train AI models.
- We don't sell your data, or hand it to advertisers.
- We don't let the app present made-up material as if it came from the books. If something falls outside what the books cover, it's built to say so.
Why this matters to us
Melis's work is about people — about thinking clearly and treating others well. It would be strange to build a tool around it that was careless with the people using it. So the app is designed to be honest about what it's doing, grounded in real source material, and respectful of your privacy at every step.
Your data is not used to train Anthropic’s models
We do not share or sell anything about you or your chats. That’s not our business.
When you use this human chat, your input is sent to our servers, cross referenced with the book content, and then processed using Anthropic's Claude API. Anthropic does not use API data to train its models, and automatically deletes inputs and outputs from its systems within 30 days. Content flagged by Anthropic's automated safety systems may be retained longer.
Links to Anthropic’s policies: Commercial retention policy, API and data retention docs, Commercial Terms.
Our formal terms and conditions are here: https://thishuman.com/terms