Knowledge and RAG

Give an agent your documents, websites, snippets and Q&A, understand how hybrid search and rerank find the right passage, and load knowledge by API.

1. 01Knowledge overviewThe four kinds of knowledge an agent can have, their limits and credit costs, and what happens when you add one.Concept2 min
2. 02Upload documentsAdd PDFs up to 50 MB to an agent's knowledge, what each page costs, and how to replace or remove a document.Guide2 min
3. 03Crawl a websiteIndex a help centre, docs site or product pages: max pages, crawl depth, include and exclude paths, statuses, re-crawls and costs.Guide2 min
4. 04Snippets and Q&A pairsAdd short facts and exact question-and-answer pairs for answers that must be word for word. Limits, costs and how they're used.Guide1 min
5. 05How retrieval works: hybrid search, fusion and rerankWhat happens between a question and the passages the model sees: pgvector and full-text search, RRF fusion, Cohere rerank, thresholds and caches.Concept2 min
6. 06Sources and citationsThe sources event in the stream, its fields, when it's sent, and how the agent cites documents and web pages in its answer.Reference1 min
7. 07Load knowledge with the ingest APIAdd documents, snippets, Q&A pairs and websites from code with /v1/ingest, and delete a source's vectors. Fields, limits, responses and errors.Reference2 min