Documents and collections
Upload files or write notes into collections or My notes. Scanned PDFs and photos are read with text recognition.
You will add knowledge — uploaded files or text you write — so it becomes searchable, either for your organization or just for you. Scanned papers and phone photos work too: their text is read automatically.
Before you start: adding to a shared collection needs Can add documents on Knowledge base for that collection. Creating and deleting collections needs Full control. My notes only need access to the app.
Create a collection
- Open Knowledge Base › Library and click New collection (with Full control).
- Type a Name, for example "HR policies", and optionally What belongs here? to help people choose where to add things.
- Check the Permission key. Roles can grant access to just this collection by its key: lowercase letters, numbers and dashes. It cannot change later.
- Save. The collection appears in the Library.
Add knowledge
- Click Add knowledge (in My notes the button is New note).
- Under Add to, pick a collection, or My notes (only you). If you can't add to any shared collection, it goes to your own notes.
- Choose How to add:
- Upload files — drop files on the box or click Choose files. PDF, Word, PowerPoint, Excel, Markdown, CSV, HTML, text, or a photo (JPG, PNG, WebP), up to 25 MB each.
- Write — a Title and the Text. Markdown is fine; headings and lists help answers cite the right part.
- Click Add (or Add N files). You see "Added — indexing now".
Everyone who can read the collection can then find it in Search, Ask and Ask Axis.
Scanned documents and photos
A scanned PDF (one with little or no text on its pages) or a photo of a page is read with text recognition while it is indexed — for example a photographed delivery note, a stamped agreement or an old printed rate card.
- Scanned PDFs up to 10 MB and photos up to 5 MB.
- Only the first 30 pages are read. A longer scan shows "Only the first 30 pages of this scanned document were read." — split it into smaller files to include the rest.
- The document shows a Read with text recognition badge. The scan is read once; re-indexing later reuses the text.
- HEIC photos (the iPhone default) are not accepted. Change the camera setting to "Most compatible", or export as JPG.
- Reading a scan counts toward your workspace's AI usage.
- Check the result on important documents: handwriting and faint stamps may be misread.
Track indexing
Each document carries a status: Indexing, Indexed (searchable) or Failed. Filter the Library by Status. Open a document to see the reason for a failure, Edit it, or Re-index it. To see when each document was last indexed and to get reminded when it is due for review, see Keep knowledge up to date.
Common failure reasons for scans:
- "No readable text was found in the scanned file."
- "The scanned file expired before it was read — upload it again."
- "Scanned document — text recognition is unavailable."
Delete
- Select documents (or Select all shown) and delete them, or delete one from its own window.
- Delete a collection from its edit window with Delete collection. Every document in it is removed from the knowledge base, and Search, Ask and Ask Axis stop using them at once.
Important: deleting a collection removes every document in it. This cannot be undone.
Tips
- Create collections by audience ("hr-policies", "sales-playbook"), because access is set per collection on the Roles page.
- Put sensitive material in a collection only the right roles can read.
Troubleshooting
"…: only PDF, Word, PowerPoint, Excel, text files and photos up to 25 MB" The file type is not supported or the file is too big.
"scanned files over 10 MB can't be read — split it into smaller files" Split the PDF and upload the parts.