📚Overview
📚 Knowledge bases give your agents searchable RAG memory. Your documents are stored, chunked, and indexed so the agent can retrieve the most relevant passages.
How they work
- Upload: Add documents such as policies, pricing, and FAQs.
- Chunking: Split them into smaller passages.
- Indexing (embeddings): Convert each passage into a representation suitable for semantic search.
- Retrieval: When a question arrives, retrieve the closest passages and pass them to the agent.
Uses
- Answer from your source instead of guessing.
- Reduce hallucination and keep responses grounded.