Single-PDF chat tools are everywhere, and they're fine when your answer lives in one file. But real questions rarely respect that boundary. "Which supplier contract has the shortest notice period?" spans five PDFs. "What do these three papers agree and disagree on?" spans three. "What changed between v2 and v3 of this spec?" spans two. To answer those, you need to chat with multiple PDFs at once — as a set, not one at a time.
This guide covers how to do that well: gathering the right documents, asking across all of them, and — the part most tools skip — checking each answer against the exact source it came from.
Why one-PDF-at-a-time doesn't cut it
Loading PDFs individually and asking each one the same question has three problems. You do the synthesis by hand, stitching answers together in your head. You lose track of which file said what. And you can't ask the questions that only make sense across documents — comparisons, contradictions, "which of these," "what's missing."
What you actually want is to treat a handful of related PDFs as one queryable set, and get a single answer that draws on all of them while still telling you where each part came from.
How to chat across multiple PDFs
1. Gather the documents behind one question
Don't try to load everything you own. Pick the specific files a question spans — the five contracts, the three papers, the two spec versions. A tight, relevant set gives better answers than a giant pile.
2. Group them into a collection
Put those PDFs into a single group so they can be queried together. In Sidenote, that's a Collection: you bundle the documents, and chat, summarise, and build a glossary across all of them at once. Collections are part of Pro, which has a 7-day trial with no card.
3. Ask in plain language
Ask the cross-document question directly — "which of these has the shortest notice period," "where do these papers disagree," "what's the combined list of requirements." The answer pulls from the whole set.
4. Verify each claim at its source
This is the step that separates a useful answer from a risky one. Every claim should point back to the specific document and passage it came from, so you can check it.
What to look for in a multi-PDF tool
| Capability | Why it matters |
|---|---|
| One answer across the whole set | You get synthesis, not five separate replies to stitch together |
| Citations name the source document | You know which file each claim came from |
| Click-to-source verification | You confirm a claim in one click instead of re-reading |
| Handles scanned PDFs | An image-only file doesn't silently drop out of the set |
| Reads in place, no upload | Fewer copies, and your existing permissions stay intact |
Sidenote is built around this: answers draw on every document in the Collection, each claim carries a citation naming its source file, and clicking it scrolls that PDF to the exact sentence. Claims that can't be grounded in a real passage are dropped, so a confident-but-unsupported answer never makes it to you. Scanned PDFs are OCR'd first, so they don't quietly fall out of the group.
Frequently asked questions
How many PDFs can I chat with at once?
That depends on the tool and plan. Sidenote's Collections hold a set of documents you group for a question — and the point is to keep the set focused, since a tight, relevant group produces sharper answers than an enormous pile. For very long individual files, see how to summarise a long PDF.
Can I tell which PDF an answer came from?
With the right tool, yes. Sidenote cites every claim by source document and lets you click through to the exact highlighted passage, so a blended answer across several PDFs always tells you which file — and which sentence — each part rests on.
Do I have to upload the PDFs somewhere?
Not with an in-browser reader. Sidenote reads the PDFs you open, read-only, and groups them into a Collection without copying them into a separate system — so there's no second set of files to manage and your existing access stays intact.