Add Your First Knowledge Document

Add the content

The current page offers two visible methods:

• Documents: select Browse Documents, search accessible PDF, DOCX, TXT, MD, or HTML files, select one, then choose Import.
• Markdown: enter Source filename, paste Markdown or use Load .md, then choose Run Pipeline.

An item marked In Knowledge already has a matching Knowledge record. Selecting it loads that record instead of creating another copy.

Prepare it for AI Agents

1. Select the new document in the left panel.
2. Under Pipeline Stages, run Markdown Parsing and inspect the result.
3. Run Structural Nodes & Chunking.
4. Run Table Summaries when the document contains tables.
5. Run LLM Enrichment when summaries, entities, PII detection, or extraction rules are needed.
6. Under Access Policy, apply the required document, page, node, or Smart Assign rules.
7. Under Embedding, choose the model and dimensions, then select Embed Document.
8. Use Testing to run similarity and nearest-neighbour checks.

Embedding makes the processed chunks searchable. Whether a user can receive them still depends on the assistant's Knowledge scope and the source access policies.