Give an agent its own knowledge base
2 min · 3 steps
Drop files into the Knowledge tab; the agent's `kb.retrieve` nodes can vector-search them.
Each agent has a per-agent knowledge base under `~/.nxr/kb/<agentId>/`. Files indexed there are vector-searchable from any `kb.retrieve` node in the agent's graph. Useful for: company handbooks the agent answers from, product spec docs, FAQ corpora, anything you want the agent to reference at run time.
Steps
- Open the agent's Knowledge tab.
Drop one or more files (txt, md, pdf, docx).
- Wait for indexing.
The KB worker chunks each file + embeds via Ollama. Progress streams as Activity events.
- Add a kb.retrieve node to your graph.
Builder canvas → drag in a Knowledge → Retrieve node. The Configure drawer lists every file you've indexed — tick the ones you want this node to read. Type plain-English instructions into the Instructions field (e.g. "Find the enterprise pricing tiers and list the monthly minimums"). The node ranks chunks against the instruction, then asks the SLM to synthesise an answer. Downstream nodes read `{{nodes.<id>.output.answer}}` for the synthesised answer, or iterate `{{nodes.<id>.output.documents}}` for the raw passages.
Live recipes need the desktop
This article is a static preview. The in-app Help sidecar inside Avery NXR can fire each step against your live project — install the desktop to use it interactively.