Knowledge Base

Create knowledge bases, load documents into them, and query them with semantic search or full retrieval-augmented question answering. Documents are chunked and embedded on write, so they are searchable the moment the call returns, and the RAG enhancements configured on a knowledge base — re-ranking, MMR, hybrid search, caching — apply automatically to every query.

📄️ List the org's knowledge bases

List every knowledge base in the organization, each with its document and

📄️ Create a knowledge base

Create an empty knowledge base to hold your own documents.

📄️ Get a knowledge base's stats + settings

Fetch one knowledge base by id, including its statistics (document and chunk

📄️ Update a knowledge base's name/description

Rename a knowledge base or change its description.

📄️ Delete a knowledge base (and its documents)

Permanently delete a knowledge base along with every document and embedding

📄️ Merge RAG settings into a knowledge base

Merge RAG settings into a knowledge base.

📄️ Semantic/hybrid search over a KB

Search a knowledge base semantically and get back the most relevant chunks

📄️ RAG question-answering over a KB

Ask a natural-language question and get a written answer grounded in the

📄️ List documents in a KB

List the documents stored in a knowledge base, with their titles, metadata and

📄️ Add a document to a KB (text or file)

Add a document to a knowledge base. The content is chunked and embedded

📄️ Get a document with full content

Fetch one document with its full extracted text, title and metadata.

📄️ Update a document (re-indexes on content change)

Update a document's `content`, `title` or `metadata`.

📄️ Delete a document (soft by default)

Remove a document from a knowledge base.