Pages cite the text taken out of each uploaded PDF, Word, Excel, PowerPoint or web file, never the file itself. Once every part of a document is verified, its original is no longer needed to answer questions.
| Choice | What happens |
|---|---|
| Keep (default) | The original stays on your deployment. Readers can download it from the source view. |
| Move to your own storage | Download the originals as a ZIP and keep them on your company drive, then remove them here. You can record where each now lives so Open the original links there. |
| Remove | The original is deleted. The text copy, every page and search stay exactly as they are. |
Rules that protect you
- Only documents with every part verified can be removed.
- Removing needs an emailed code tied to exactly the files chosen, and cannot run during a build.
- Removing a file never changes who can see its pages: the original’s folder still decides its team.
What you give up
Reading a removed file again, for example with a better chart reader, needs it uploaded again, and its pages are not rebuilt automatically when the AI writer changes. Uploading the same file again restores it.
Find it on the console Team screen, under Storage for original files.
