Search across PDFs
A name, an amount, a date: found across hundreds of PDFs, with nothing ever uploaded.
Your files never leave your deviceDrop a whole folder of PDFs: Vellum reads the text of every page and files it into an index on this device, then you search a name, an amount or a date across every document at once.
Drop your PDFs here
or drag many files at once
Your cabinet
The index stays in this browser, on this device, until you clear it. Nothing is ever sent.
Frequently asked questions
Are the PDFs themselves sent anywhere to be indexed?
No. Each page is read by your browser (pdf.js), and only the text that comes out of it is saved in IndexedDB, a database built into the browser. The PDF file is never transmitted, nor even copied elsewhere on your device.
Why doesn’t a result’s preview always show up?
Because the index keeps only the text, not the file — that is what keeps it tiny even facing a folder of several hundred PDFs. As long as the file is still open in the current tab, its preview shows; after reloading the page, it needs to be dropped again to see it.
Does the search care about case and accents?
No, on purpose: searching for “geneve” also finds “Genève”, “GENÈVE” or “genève”. Several words in the same search must all appear on the same page, in any order.
Does an amount like 1’200 show up when I just type 1200?
Yes. The search strips the Swiss apostrophe and non-breaking space that separate thousands before comparing, so “1’200”, “1 200” and “1200” all find each other — both in the document text and in what you type.
What happens with a scanned PDF that has no text?
A page with no detected text is counted and reported in the summary after indexing — it doesn’t cause an error, but it stays unfindable by search until it has been run through the text recognition tool, which adds it a text layer.