Use this knowledge base from a chat application, coding agent, or your own software. Every export comes from the same pages and datasets as the site. Start with a subject pack for a chat, connect an agent for repeated questions, or download the complete searchable database for offline work.
The subject directory lists small plain-text packs grouped by category. Open a pack or upload it to your chat application. Each passage includes its page identifier, canonical address, heading, and existing reference links. Large subjects span several numbered parts; read all parts when completeness matters.
For a focused question, open a cortisol feedback context bundle. Change the query to your subject. The response includes selected passages and references, with an explicit truncation flag. A focused bundle is a selection, not an exhaustive account of everything in the database.
--db with your database filename.Configure your application's remote MCP connection with https://teahann.com/mcp. The read-only tools are search, get, batch, neighbors, and context. Connection support depends on your application and account. No API key is required for this public corpus.
The HTTP API provides the same operations under /api/v1/. Try search, a page record, or its outgoing links. Use max_chars to bound context responses; this measures serialized characters, not model tokens.
Identifiers describe pages. Two domain pages about the same person remain distinct unless explicitly reconciled. Relationships distinguish hyperlinks, authored related-item lists, and directory membership. A link does not establish causation, agreement, or a scientific mechanism. Citations are inherited from the site and are generally page-level references; export does not independently verify the claims.
Original units, qualifiers, and dataset provenance are retained. Read Sources & Licences and the exported provenance records before redistributing material. The export does not grant a new licence over its source datasets.