How RAGBot works — end-to-end
This is the daily loop in the UI.
Step by step
A. Get in
Sign in → marketplace → RAGBot. Check Dashboard so you know if anything is already loaded.
B. Prepare knowledge
- Knowledge Repositories → + New Repository.
- Documents → pick that repository → Choose File → Upload & Index.
- Wait until Chunks is greater than 0.
C. Ask
Chat Workspace → select the same repository → type a question. Keep questions about the files you uploaded (names, IDs, facts in the PDF/DOCX/CSV).
D. Revisit
Conversations to find an old Q&A. Reports to see storage and who chatted. Audit Logs if you need “who called which API”.
E. Clean up
Trash icon on a document or repository removes that content. Dashboard and Reports counts update after that.
What “chunks” and “vectors” mean
| Word | Meaning for you |
|---|---|
| Document | The file you uploaded (green_dc.pdf) |
| Chunks | Pieces of text the search engine uses |
| Vectors | Numeric fingerprints of those chunks |
| Repository | A folder of documents you chat against |
If chunks stay at 0, indexing failed — try another format or a smaller file. See Troubleshooting.