Built-in RAG search
Built-in RAG search means Synamcps embeds ingested knowledge and serves semantic queries over REST (POST /api/knowledge/search) and MCP knowledge tools — filtered by ACL, storage, and source labels.
Semantic retrieval that respects storages, sources, and document visibility.
pgvector or Qdrant
Choose the vector backend that fits your stack. Embeddings are produced during ingestion so search stays fast; filters for source, sourceUrl, and storageIds keep queries scoped to the right corpus.
REST and MCP retrieval
Call POST /api/knowledge/search from pipelines, or let agents use knowledge_* MCP tools. Either path applies the caller's token scopes and document visibility — retrieval never bypasses governance.