Frequently asked questions
Does this work with scanned PDFs?
Not yet. Scanned PDFs are photographs of paper with no text inside — they need OCR, which is on our roadmap. This tool works with digital PDFs: if you can select text in a PDF viewer, it will extract.
Is it safe for bank statements?
Safer than any upload-based converter, because there is no upload. Extraction runs in your browser's memory and nothing is sent to us — we couldn't see your statement if we wanted to.
The columns came out slightly wrong — why?
PDFs don't actually contain tables, just positioned text, so reconstruction is an educated guess. Complex layouts (merged cells, multi-line rows) may need a quick tidy-up — the CSV cleaner helps.
What's the difference between the Excel and CSV downloads?
Excel (.xlsx) keeps each PDF page on its own sheet. CSV joins all pages into one flat file — better for importing into other software.
Turning a PDF back into data
A PDF looks like it contains a table, but underneath it's just hundreds of pieces of text, each pinned to an x/y coordinate on the page. There are no real rows or columns — your eye assembles them from the spacing. To rebuild a spreadsheet, this tool reads every text fragment and its position, then clusters fragments that share a vertical line into rows and fragments that share a horizontal band into columns. It's the same coordinate-clustering technique the expensive desktop converters use, running locally in your browser.
What it handles well — and what it doesn't
Clean, digital PDFs with regular tables — bank statements, invoices, price lists, exported reports — convert accurately. Where reconstruction gets harder is messy real-world layouts: merged cells, rows that wrap onto two lines, or columns separated by nothing but a hair of whitespace. In those cases the structure is an educated guess and may need a quick tidy-up afterwards. A useful pattern is to convert here and then run the result through the CSV cleaner to drop stray blank rows and normalise the columns.
One hard limit worth knowing: this tool cannot read scanned PDFs. A scan is a photograph of paper — there is no text inside it to extract, only pixels. Pulling data out of a scan requires OCR (optical character recognition), which is on the roadmap but not available yet. The quick test: if you can select the text with your cursor in a normal PDF viewer, this tool can extract it; if you can't, it's a scan.
Why do it in the browser?
Bank statements are the single most common thing people convert to Excel, and they're also among the most sensitive documents you own. Every upload-based converter asks you to send that statement to a stranger's server. FileKitt doesn't — extraction runs in your browser's memory and nothing leaves your machine, so the privacy question simply doesn't arise. You get a genuine .xlsx workbook (one sheet per page) or a single flat CSV, and you can preview the parse before downloading either.
Getting the cleanest export
- Use the original digital PDF, not a printed-and-rescanned copy — the scan will have no extractable text.
- Pick the right download. Choose Excel to keep each page on its own sheet; choose CSV to get one continuous file for importing into another system.
- Expect to tidy complex tables. For anything with merged headers or multi-line entries, budget a minute to fix a column or two — it's still far faster than retyping.