Scanned PDF — two jobs, not one

The job behind the search

People type “scanned pdf to word” because they want to keep working. They have a phone photo of a letter, a faxed form, or a paper invoice that was “printed to PDF” as an image. They need to copy a sentence, search for a name, or send the file into an editor or a spreadsheet.

The search names Word because Word is where they expect to type. The blocker is earlier. A scan has no text layer. A converter that only wraps the image in a .docx has not produced an editable document. It has produced a photocopy with a Word extension.

What Papermeld actually produces

The conversion that ran on a real scan today produces a PDF with a text layer. You can search it. You can select a line and copy it. The original page image stays on the page. That is the file we will hand back when the first job succeeds.

The independent check is simple and strict. The result has to be a PDF, and it has to contain readable characters in its text layer. If that layer is missing, the original scan is not delivered as if it were searchable. You get a sentence in the conversation instead.

This is not a promise that every character is right. Recognition can glue a digit to a letter or drop a space. On one invoice we ran, a quantity that read “42 QTY” on the page came back as “420TY”. The file was still searchable. It was not a verbatim transcript. Papermeld will not claim character-level accuracy, and this page does not state an accuracy percentage.

The second job is separate

An editable Word file, or a spreadsheet you can sort, still needs a text layer to start from. The PDF to Word conversion on this site refuses a photocopy: if the words live only in an image, that job is not treated as finished. PDF to Excel will not turn a picture of a table into cells.

So the honest sequence for a scan is: make the PDF searchable first, review the text against the page, then ask for Word or Excel if you still need an editor or a grid. Do not expect this step itself to hand you a structured Word outline or a workbook with headings and tables already rebuilt.

If you already have a text-based PDF, skip this page. Word is the PDF to Word conversion. A table-dense report is the PDF to Excel conversion. Those are different jobs, with their own failure modes.

What you should still review

Open the searchable PDF. Search for a word you can see on the page. Select a line. If you cannot select text, the job did not finish.

Then read the numbers that matter. Quantities, dates, names, and totals are where recognition errors concentrate. Correct those against the scan before you paste them into Word, Excel, or anything else.

You remain responsible for the text you use. Papermeld will not present a guessed page as a finished Word file or a finished spreadsheet.

How to start

1. Upload. Sign in, land in this product’s conversation, and upload the source file there. There is no separate form on this site.
2. Say what you want. Tell the conversation this PDF is a scan and you need a searchable copy — text you can select and search. If you later need Word or Excel, say that only after the scan has a text layer.
3. Wait. Wait while the conversion runs. A file is offered only after an independent check of the output has passed. A failed job is a sentence in the conversation, not a download.
4. Download. Download the file and review it against the source before you send it. You remain responsible for that look.