Diff Checker

Paste the first version into the Original box and the second version into the Changed box. The comparison appears under the boxes straight away and updates as you type, so you can fix a line and watch its colour change. You can also drop a file on either box, or press File above it to open one.

This diff checker online works in any modern browser on a computer, tablet or phone. Both texts stay on your own device: the comparison is done by code in the page, and nothing is sent to a server. There is no account to make and no limit on how often you use it.

Lines that were removed are shown in red, lines that were added are shown in green, and lines that were edited are shown as a pair, with the exact words or characters that differ marked in a darker shade. A summary line counts the changes, and the arrow buttons jump from one change to the next.

The bar above the result gives the summary of changes. It reads, for example, 4 changes, followed by +3 added, −1 removed and 2 changed lines. A change is one block of lines that differ, with matching lines on both sides of it. One block can hold several lines, so the number of changes is often smaller than the number of lines.

An added line exists only in Changed and is shown in green. A removed line exists only in Original and is shown in red. A changed line is a line in Original that was replaced by a line in Changed: the tool shows the two side by side, the old one on red and the new one on green, and marks the exact part that differs in a stronger colour. In Side by side view, the empty space opposite an added or removed line is light grey, so the two columns stay lined up.

Here is a short example. Original holds three lines: Apples 3, Pears 5 and Plums 2. Changed holds Apples 3, Pears 6, Plums 2 and Cherries 4. The first and third lines match and have no colour. Pears 5 and Pears 6 are paired as a changed line, with only the number marked. Cherries 4 is a green added line at the end. The bar reads 2 changes, +1 added, −0 removed and 1 changed lines, because the edited line and the new line are two separate blocks.

When the two texts match, the bar says The two texts are identical. If any Ignore option is ticked, it adds (with the options above), as a reminder that small differences may exist but were set aside on purpose.

Most people need a diff checker for text that a computer did not write: a contract that came back from the other side, an essay after a round of edits, a policy page, a translation, or a list of names. Reading two versions side by side by eye is slow and it is easy to miss one changed word in a long paragraph. Here each edited line is paired with its new version and the edited words are marked, so the eye goes straight to them.

The comparison works line by line, and for prose a line is usually a whole paragraph. That is fine: the words that changed inside the paragraph are still marked. If a paragraph was split in two, you will see the old paragraph as changed and the second half as an added line.

If the texts were copied from different places, small differences in spaces or in capital letters can hide the real edits. Tick Ignore spaces and Ignore case to set them aside, and Ignore blank lines if one copy has empty lines between paragraphs and the other does not. To count the words in either version, paste it into the Word Counter; to change its letter case first, use the Case Converter.

Mark inside lines decides how finely an edited line is marked. With Words, the line is split into words, numbers, single punctuation marks and runs of spaces, and the parts that differ are marked as whole words. This is the easiest way to read changes in sentences.

A diff checker by character goes one step further: with Characters chosen, every letter, digit and symbol is compared on its own. That is the right choice for things where a single character matters, such as a version number, a price, an ID, a URL, a date, a password hint or a line of code where only one bracket moved. For example, comparing order-1045 with order-1054 by Words marks all of 1045 and 1054, while Characters marks just the one or two digits that moved.

The choice only changes the marking inside lines. Which lines count as added, removed or changed stays the same. For a very long line that is completely different on both sides, the tool marks the whole line instead of hunting for tiny matches, which keeps the page fast.

Spaces are the most common source of noise in a comparison. One editor turns tabs into spaces, another strips spaces at the end of lines, and a copy from a web page adds a space here and there. Without help, every one of those lines shows up as changed even though nothing that matters is different.

A diff checker without whitespace noise is what Ignore spaces gives you. With it ticked, any run of spaces or tabs counts as a single space, and spaces at the start and end of each line are ignored. So a line indented with a tab matches the same line indented with four spaces, and two spaces between words match one. Lines that are the same apart from this are shown as unchanged.

Note what it does not do: it does not remove spaces altogether, so a + b and a+b are still different. Ignore case works the same way for capital letters, and Ignore blank lines leaves out empty lines on both sides, which helps when one version has extra line breaks between paragraphs. The three options can be combined.

Two JSON files can hold exactly the same data and still look completely different as text. One may be on a single line, the other indented; one may list name before id, the other id before name. A plain line diff then marks almost everything.

That is why a diff checker json mode is built in. Tick JSON (sort keys) and both texts are read as JSON, the keys of every object at every level are put in alphabetical order, and both are written out with a 2-space indent before they are compared. Key order and layout stop counting, and only real changes in values, added keys and removed keys remain. Arrays keep their order, because in JSON the order of array items is part of the data.

If a box does not hold valid JSON, that side is compared as plain text instead, so nothing is lost. To find the error, paste the text into the JSON Formatter, which shows the line and column of the problem and can repair common slips such as trailing commas. When you need the data in another shape, CSV to JSON and JSON to CSV convert between the two.

A diff checker for code is useful whenever two versions of a file are not in the same repository: a snippet from a forum against your own copy, a config file from a server against the one on your laptop, or two versions of a script that someone sent by email. The file picker reads common code and data files directly, including JS, TS, TSX, JSX, Python, Java, C, C++, C#, Go, Ruby, PHP, SQL, CSS, HTML, XML, YAML, INI and log files.

For code review, the Side by side view with line numbers works like the review screens people know from version control. Characters is a good marking choice for code, because a changed operator or a missing semicolon is a single character. Ignore spaces hides indentation and formatting changes when you only care about logic; untick it again when indentation matters, as it does in Python and YAML.

Only changes keeps each change with 3 lines of context around it, the same amount most review tools show by default. When you are done, press .diff to save the comparison and attach it to a ticket or a message.

A .diff file is plain text in the unified style that Git and other version control tools use to show changes. It starts with two header lines: --- original names the old version and +++ changed names the new one. After that, every line starts with one mark: a space for a line that is the same in both, a minus sign for a line that was removed, and a plus sign for a line that was added. An edited line appears twice, first with a minus sign and then with a plus sign.

The file written here contains every line of the comparison, not only the changed parts, so anyone can read the full context in any text editor. It does not include the @@ hunk headers with line ranges that patch tools need, so treat it as a readable record of the changes to share or keep, not as a patch to apply with a command.

The .diff file follows the options you chose. With Ignore blank lines ticked, empty lines are left out of it too, and with JSON (sort keys) ticked, it holds the sorted, formatted JSON. Only changes affects the screen only: the download always holds the whole comparison.

You can use this as a diff checker for PDF text. Drop a PDF on either box and the text of every page is read in your browser and placed in the box, page after page, with line breaks where the PDF has them. Then it is compared like any other text. This is handy for checking what changed in the wording of a contract, a report or a form between two versions.

Only the text layer is read. A scanned PDF has no text layer, so the box stays nearly empty; run it through Image to Text first to get the words out. The way a PDF stores its text also means line breaks can fall in slightly different places in two versions of the same document, which can show up as changed lines; Ignore spaces and Ignore blank lines help with that.

When the layout matters, such as moved images, changed fonts, a new logo or a table that shifted, a text comparison is not enough. Compare PDF shows the two documents page by page with the changes highlighted on the pages themselves, including a visual comparison. To keep the text of a PDF for editing, PDF to Text and PDF to Word save it as a file.

A diff checker for Excel files saves you from scanning two spreadsheets cell by cell. Open an .xlsx file in a box and each row becomes one line, with the cells of the row separated by tabs. Cells are read as they are displayed, and dates are written as year-month-day. If the workbook has more than one sheet, every sheet is included, each starting with a line that holds a # and the sheet name.

Because each row is a line, an added row shows up as an added line and an edited cell shows up as a changed line with the new value marked. This works best when both files have their rows in the same order; sort both the same way first if they do not. CSV and TSV files are compared the same way, one row per line, without any conversion.

Word .docx files are read as plain text, one paragraph per line. Formatting such as bold, colours and fonts is not compared, only the words. Older .xls and .doc files are not supported; save them as .xlsx or .docx first. To turn a spreadsheet into a document for sharing, Excel to PDF keeps the layout, and PDF to Excel brings tables from a PDF back into a sheet.

There is no fixed size limit, because the files never leave your device: what your browser can hold, it can compare. A diff checker for large files has two jobs, though: finding the changes quickly, and showing them without freezing the page. Both are handled.

The comparison uses the Myers algorithm, which does work in proportion to the number of differences, not only the size of the files. Two long log files with a few changed lines are compared very quickly. If two texts differ in more than about 6,000 places, a full search would take too long, so the tool switches to a simpler mode: line 1 is compared with line 1, line 2 with line 2, and so on. The bar then says Very different texts: compared line by line in order.

The result shows up to 6,000 lines at a time. For bigger files, the note under the result tells you so; tick Only changes to see the changes with 3 lines around each, which is usually all you need. The .diff download always holds the whole comparison. Typing in a huge box repeats the comparison after each change, so for very big texts it is quicker to paste the final version than to edit it in place.

Each text is split into lines. Windows and old Mac line endings are turned into plain line breaks first, so a file saved on Windows and the same file saved on a Mac match. Then any options you ticked are applied to a copy of each line used only for matching: spaces collapsed, letters made small, blank lines skipped. The original lines are still what you see.

Next, the Myers algorithm finds the shortest list of additions and removals that turns Original into Changed. It is the same family of algorithm behind the diff command and Git, and it keeps as many lines in place as possible, so a line inserted near the top does not make everything below it look different.

Because the matching copy is used only to decide which lines are equal, the text you see is never altered. With Ignore case ticked, for example, Hello and HELLO count as the same line, but each side still shows its own spelling, so you can see what the original really said.

Where a block of removed lines sits right next to a block of added lines, the tool pairs them up one by one as changed lines, and runs the same algorithm again inside each pair, on words or characters, to find the exact edits. Any extra lines in the block stay as plain added or removed lines.

This is a diff checker free for anyone to use, as often as needed. There is no sign-up, no account, no daily limit and no paid version: every option on the page is open to everyone.

Both texts, and any files you open, are read and compared by code running in this page on your own device. Nothing is uploaded or stored on a server, which matters when you compare contracts, private notes, customer lists or source code. Only your view and option choices are remembered in your browser; the texts are gone when you close the tab.

Once the page has loaded, comparing keeps working if your connection drops. Opening a PDF, Word or Excel file for the first time loads a small reader in the background, so do that while you are still online.

Comparing PDFs where layout matters? Compare PDF shows changes on the pages themselves. Need to tidy JSON before comparing? The JSON Formatter checks, repairs and sorts it. Want to prove two files are byte-for-byte the same without reading them? The Hash Generator gives each one a fingerprint you can compare.