Convex Agents RAG

To add retrieval-augmented generation to Convex Agents, integrate the rag module in convex.config.ts and ingest your documents into separate namespaces for semantic search. This grounds agent responses with knowledge-base context.

Semantic search over namespaces is a retrieval mechanism that allows AI agents to query specific document sets, such as support docs or policies, to provide accurate and context-grounded answers.

Ingest documents into a configured namespace to populate the knowledge base. Once ingested, your Convex Agent can leverage the rag module to query this content and answer with sourced context.

Yes, you can ingest content into separate namespaces. This allows your agents to perform semantic search over isolated document sets, ensuring responses are grounded in the correct policy or product data context.

Yes, retrieval-augmented generation supports searching product FAQs and support docs. The skill grounds agent responses by applying semantic search over your ingested knowledge base content.

Agent responses lack knowledge base context if documents are not properly ingested into namespaces or the rag module is not configured in convex.config.ts. Grounding requires semantic search integration to source relevant information.