Choose the two files
PDF or XLSX up to 20 MB. CSV or TSV up to 2 MB.
Choose the second copy in the same supported format.
PDF: 200 pages. XLSX: 50 sheets/100k cells. CSV/TSV: 10k rows/256 columns.
Document bytes stay in this browser tab. Downloaded reports contain extracted text and filenames. Check the report before sharing it.
Compare the Two Versions
Limits: 200 pages/PDF; 50 sheets, 100,000 cells/XLSX; 10,000 rows, 256 columns/CSV or TSV.
This PDF comparison tool compares readable text and page dimensions. The page inventory mode checks page counts and dimensions by page number. Neither mode compares images pixel by pixel.
Content pairing uses matching readable pages as anchors. An inserted page can then appear as an addition while later pages retain their original references. Unmatched gaps are paired in order.
Page-number pairing compares original page 1 with comparison page 1 and continues in that order. Use it when you know the copies have matching pagination. Duplicate pages can make content pairing ambiguous.
Whitespace normalisation collapses spaces and line breaks. It preserves case, punctuation and digits. Switch it off to inspect extraction spacing. Text order changes can still produce a reported difference.
Choose files, set the mode and select Compare files. After changing a file or setting, compare again. PDF parsing starts only when you select the button.
Inspect a Reported Difference
Open a reported difference to see the original and comparison file names with their page numbers. Changed text items retain their extracted x and y coordinates in PDF points.
The coordinates use the PDF’s page coordinate system. Open the named page in your PDF reader to check the surrounding transaction. A rotated page can use a different coordinate orientation.
For example, changing “Paid $100” to “Paid $900” reports changed text and a numeric-text difference. The numbers flag also catches dates and reference numbers. It does not classify every number as money.
A line break or repeated space can disappear under whitespace normalisation. A changed digit remains. If the extracted items are unchanged but appear in a different order, the report retains the full page text for inspection.
Download the JSON report to retain complete source references and extracted text items. Keep both original files with it. The report does not include the original PDF bytes.
Understand the Comparison Limits
A scan without a text layer is marked unreadable in text mode. This tool does not perform optical character recognition. Blank pages also need manual inspection.
Different fonts or extraction order can change extracted text without changing the meaning. Matching text can hide different images, signatures or annotations. Inspect those features in the source documents.
Encrypted, damaged or unsupported files stop the comparison. A PDF with an incomplete text layer can still omit visible content. Page inventory mode checks dimensions and count only.
A moved page is an exact text match at a different position in an equal-length document. Added or removed pages can shift other page numbers. The report preserves those pairings without claiming each shift is a separate move.
A difference alone cannot prove fabrication or identify the authorised version. Use the source records and the checks for a suspicious bank statement when assessing authenticity.
Compare Spreadsheet Versions
Select spreadsheet mode for two XLSX workbooks or two UTF-8 CSV/TSV exports. XLSX comparison pairs worksheets by their exact names and cells by their addresses. Hidden worksheet cells are included.
Each changed value or stored formula has a sheet and cell reference, such as Transactions!D12. Added or removed worksheets also appear. Inserted rows can produce many cell differences because addresses remain fixed.
XLSX formulas are compared as stored text with their attributes. Shared formulas retain their shared-definition references. The tool compares cached values separately and does not recalculate formulas.
External relationship targets in the supplied XLSX are listed with their package part references. They can include linked workbooks or hyperlinks. The tool displays these targets as text and never opens them.
CSV and TSV have one sheet, labelled Sheet1 here. Formula text and external reference candidates appear only if the export contains them. Ordinary value-only exports lose workbook formulas and relationship records.
XLS, XLSB, ODS and password-protected workbooks are unsupported. Charts, formatting, named ranges and macros are outside the comparison. Chart-only sheets stop XLSX processing. Stored numeric values can be date serials.
The spreadsheet schedule records changes for review. A corrected formula or authorised update can be legitimate. Check the original linked sources before sharing a workbook or treating a change as improper.