Feed Janore your knowledge.

Source types

Add as many sources as you like. They all ingest in parallel — most assistants are fully indexed in a few minutes.

Website

Paste any public URL. Janore tries your sitemap.xml first, then falls back to a same-origin BFS of the homepage. Up to 25 pages per source. Each page is split into ~800-token chunks tagged with its URL so Janore can cite the right page.

FAQ

Paste a Q/A block, or pick a starter template (Restaurant, Real estate, Medical clinic, E-commerce, SaaS, Hotel, Beauty / Salon, Freelance, Generic). Edit the template until it sounds like you.

Notes

Free-form text. Use this for anything that doesn't fit as Q/A — your address, founding story, return policy, sample answers in your voice.

Files (PDF, DOCX, TXT)

Upload up to 10 documents per source, 20 MB each. Text and tables are extracted, scanned PDFs are OCR'd. The original file is kept; replies cite the document name and page when relevant.

How ingestion works

Each source is enqueued for processing. URL crawls run in parallel; FAQ, notes and files are chunked and embedded by a 5-minute background job. You'll see live status (pending → ingesting → ready) in the Knowledge tab.

Re-ingestion

Edit, re-crawl or delete every source post-creation from the Knowledge tab of your assistant. Edits trigger a re-embed within 5 minutes.

Privacy & RGPD

Knowledge content is treated as your data: stored in the EU, never used to train shared models, deletable on demand. The Account → Data export endpoint includes every source you've added.