PDF to CSV
Convert PDF tables into structured CSV format.
Drag & drop your PDF here
or click to select file
User Guide
Check the table is real
Try selecting a row in your PDF reader. If the text highlights, the table is made of characters and can be converted. If nothing selects, you have a picture of a table, and no converter will read it until OCR has turned those pixels into text.
Upload and convert
Add the PDF — one file, up to 10 MB — and press Convert to CSV. The document is processed on a server and the result is offered for download.
Open it in a spreadsheet
Import rather than double-click, and tell the import dialogue that the separator is a comma. Double-clicking hands the file to whatever defaults your spreadsheet has, which is how leading zeros vanish from account numbers and how dates silently change format.
Check the seams
Scan the first rows, the last rows, and any row where a cell wrapped onto two lines in the original. Those three places are where conversion errors concentrate. Repeated page headers appearing mid-file are common in multi-page tables and are quick to delete once spotted.
About the PDF to CSV Converter
Statements, price lists, timetables and exported reports arrive as PDFs, and all of them are tables somebody eventually needs in a spreadsheet. Retyping is slow and puts errors exactly where errors are expensive.
How a table is found in a file that has no tables
This is the part worth understanding. A PDF has no concept of a table, a row or a cell. It has glyphs at coordinates. What the converter does is reconstruct the grid by inference: it looks for ruled lines, for runs of text sharing a vertical position, and for consistent horizontal alignment across many lines, then decides where the column boundaries must be.
That works well when the visual cues are strong — ruled borders, one line per row, consistent alignment. It struggles precisely where a table is designed to look elegant: merged header cells spanning two columns, a total row set apart by whitespace, or a cell whose text wraps onto a second line. Each of those carries meaning to your eye that has no equivalent in coordinates, so the converter must guess, and it will not always guess as you would.
The one limitation to plan around
When a document yields several separate tables, the current version returns the first. If you need a specific table from a long report, split those pages out first — the Split PDF tool does that in your browser at no cost — and convert just them.
Why import rather than double-click
RFC 4180 describes CSV as plain text with commas and quoting rules; it carries no types at all. Your spreadsheet therefore guesses what each column is, which is how leading zeros vanish from account numbers and how 03-04 silently becomes a date. Use Data then From Text/CSV and set the awkward columns to Text in the preview. Double-clicking hands those guesses free rein.
Check the seams
Scan the first rows, the last rows, and any row that looked tall in the original. Conversion errors concentrate there. In a multi-page table the repeated column headers usually arrive as ordinary data rows and are quick to filter out once you know to look.
Frequently Asked Questions
My PDF has several tables but I only got one.
That is the current behaviour — when the conversion yields multiple tables, the first is returned. Split out the pages holding the table you want and convert those on their own. Splitting runs in your browser, so it costs nothing and adds no upload.
Why are some rows merged together?
Usually because a cell’s text wrapped onto a second line in the original. Visually it is one row; structurally it is two lines that the converter has to decide about. Check any row that looked tall in the PDF.
My numbers turned into dates in Excel.
That is Excel, not the conversion. Use Data then From Text/CSV, and set the offending columns to Text in the import preview. Opening a CSV by double-clicking lets the default guesses run unchallenged, which is where 03-04 becomes a date.
Can I convert a scanned bank statement?
Not directly. A scan contains no characters, so there is nothing to place into cells. Run OCR over it first to create a text layer, then convert the result — and check that conversion carefully, because OCR errors in digits are easy to miss.
Is there a page limit?
No page limit, but a 10 MB file limit. Long statements usually fit comfortably; if yours does not, split it into parts and convert each.
Which is better for me, CSV or XLS?
CSV if anything is going to read the file programmatically, or if you want something that will open anywhere in ten years. XLS if you simply want to double-click it into a spreadsheet and start working, and you do not need the file to travel.