Just tried to fix a typo in a source document I added to NotebookLM. I can't edit it in the app—I have to go back to the original doc, make the change, and re-upload. It feels like a weird oversight for a tool built around source material.
This is a real workflow killer. If I'm summarizing or asking questions about a project spec, and I spot a small error, I want to correct it right there. Now it's a whole detour. Anyone else running into this?
You're definitely not alone. I see this come up a lot in the feature requests channel.
The workaround is annoying, but from a compliance standpoint, I get why they're hesitant. It creates two versions of a "source" document - one in your drive and one in the app. That can be a real problem for audits or maintaining a single source of truth.
Still, for personal use, it's a friction point that shouldn't be there. I'd love a toggle that lets you enable direct edits for non-critical documents, with a clear warning about version drift.
Review first, buy later.
Yeah, the compliance angle is a real blocker, but I think it's solvable. You could treat it like a code review system - any edit made in the app creates a "patch" or a suggested change file that you can then apply back to the original source. It wouldn't be a silent fork.
That toggle idea is good, but I'd want it to be more granular, like document-level permissions. Having a big "enable source editing" flag feels too all-or-nothing. Honestly, the bigger issue might be the mental model - if I'm asking the AI questions about a doc, and I can tweak the text, does it re-process the whole thing? That could get weird fast.
pipeline all the things