PDF Text and Metadata Inspector
Choose the pages whose text you need and read the metadata the file declares. Up to 30 pages are read per run in this browser, and the result copies out as plain text.
Runs locally in your browser ·
Choose the pages whose text you need and read the metadata the file declares. Up to 30 pages are read per run in this browser, and the result copies out as plain text.
Runs locally in your browserOpen a PDF, name the pages you need and press the run button: the output prints the Title, Author, Subject, Creator, Producer and page count the file declares, then the text layer of every selected page under its own --- Page N --- heading. The file never leaves the browser, and the reading is done by the bundled pdf-lib and pdf.js libraries.
A three-page document with a real text layer came back in about 0.3 s, and a 52-page fixture answered the 30-page cap with a written message until the range was narrowed to 1-30. A page without a text layer - a scan, for example - returns no characters, and the tool says that instead of printing an empty result.
Metadata: the Title, Author, Subject, Creator and Producer entries plus the page count, exactly as the file declares them. A document written by pdf-lib names pdf-lib as its producer even when it has no author, and every field the file leaves out is printed as an em dash.
Text: the text layer of the selected pages, read with pdf.js. The extraction keeps the line breaks of the page, joins the fragments that share a line with a single space, and prints pages in the order you listed them, so 3-1 shows page 3 before page 2. Nothing else is interpreted - images, form fields and annotations stay untouched.
Pages is the page count of the whole document while Pages to inspect counts only this run: the 52-page fixture with 1-30 reports 52 and 30. Input size is the size of the file you opened.
The 30-page cap exists because every page is parsed in the browser. The first run of a session also loads the PDF engine, which takes a few seconds on a cold cache; the 30-page run itself finished in about 0.3-0.4 s, and a blank field on a longer document stops with the cap message - narrow the range and run again.
One PDF per run, up to 250 MB, and no OCR: a scanned page carries a picture of the text, not the text, so the tool reports that no text layer was found. Encrypted or password-protected files are refused, and a damaged file answers This file could not be read as PDF. Check that it is not damaged. instead of a raw parser error.
Everything happens in the page: the PDF is not uploaded, the tool keeps working offline once loaded, and the result is plain text you can paste anywhere. The inspector only reads and copies - metadata and pages are never written back into the file.