AI that knows your business — without a shared prompt dump.
Connect documents to retrieval, agents, chat and workflows. Knowledge stays scoped to the organization using it, so another organization does not see your context.
Knowledge → Index → Retrieve → Agent → Action
What RAG actually does
Core knowledge capability. Binary file storage is a separate licensed files add-on.
Document ingest
Tenant documents are ingested into a text knowledge base used by website chat and workflow retrieval steps.
Hybrid retrieval
pgvector similarity is combined with lexical ranking so keyword and semantic matches both count.
Tenant filter
Retrieval is scoped to the tenant in the JWT. Another organisation does not see your chunks.
Website chat
Licensed website chat sits on the same RAG layer: widget answers and handoff into other graphs.
Not the file store
Photos and PDFs kept as files live in the files add-on (MinIO). RAG is text knowledge, not blob storage.
Ready to move beyond the chatbot?
Build AI that understands your business, follows your processes and can actually get things done.