Sebastian is an AI
Sebastian, the guide in Seba’s dream, tarot, I Ching and astrology readings and in its chat, is an AI. It is not a person and not a therapist, and it has no biography, face or credentials. Its answers are written live by large language models: DeepSeek V4 for chat, dreams, tarot, astrology and research, and OpenAI’s GPT-5.6 for the I Ching. Smaller models (DeepSeek V4 Flash and Anthropic’s Claude Haiku) help with search. No person reads a reading before you do.
How a reading stays tied to the books
Before Sebastian writes, Seba’s own code, not the model, searches the library and hands it a short packet of passages, each with its source. Sebastian quotes from that packet. Code then checks every quotation against the passage it names, and a quotation that does not match is removed before you see the reading. The sources a reading drew on are listed under it and link to their pages in the library.
When someone may be in danger
Seba is not therapy and not a crisis service. If what you write suggests you may be in danger, the reading pauses and Seba shows crisis lines, such as 988 in the United States. To talk with a person trained in depth work, the directory lists analysts and practitioners.
Pages written with AI
Library commentaries are written with AI from the full text of each book, under the editor’s direction. Since October 2026 every quotation in a new library page is checked word for word against the book before the page goes up.
The concordance and the short concept entries are drafted with AI from passages in Seba’s library. They are reference aids that point back to the books they quote.
Directory listings come from public sources, such as training institutes’ rosters and the professionals’ own sites, and each record keeps its source. The short descriptions are written with AI from those sources. Anyone listed can correct their listing or have it removed, and a removal is permanent. A phone number shows only when the person asked for it, and email addresses are never shown.
What a person decides
Cody Peterson, Seba’s founder and editor, decides what Seba covers and what gets built, writes the rules its AI writers work under, and answers corrections himself. He builds and runs the site with AI coding agents (Anthropic’s Claude Code and OpenAI’s Codex) and checks that what they make works.
Seba does not invent people, reviews, photos or credentials, and it does not present AI writing as anyone’s personal account.
Your words
What you type to Sebastian is sent to the AI provider that writes the answer, under that provider’s terms for its API. If you speak instead, Cloudflare Workers AI turns your voice into text, Seba does not save the recording, and you see the text before anything is sent. The privacy policy has the details.
Found a mistake?
Write to cody@seba.health with the page and, if you have one, the source that shows the error.