Code-Diff - Confronto di codice

Something changed, but what exactly? That question comes up constantly: the landlord sends a new version of the lease, the tax office a corrected schedule, a colleague a reworked config file. Comparing line by line by hand is tedious and error-prone - you always miss the one small clause that matters. Code Diff makes the differences visible so nothing slips past you.

Under the hood it runs the Myers diff algorithm, the same approach git uses. It finds the shortest sequence of edits between two texts instead of naively pitting line against line. The result is a clean diff that does not mark everything after a single inserted blank line as changed. Moved blocks show up as removed in one place and added in the other.

You can switch between two views. Side-by-side shows old on the left, new on the right, with synchronised scrolling. The unified view stacks them, the way diffs look in a terminal. Changed lines also get word or character level highlighting, so you see not just that a line changed but which word. A minimap on the edge shows every change at a glance and jumps straight to it on click.

Unimportant differences can be hidden. Ignore whitespace, case, or blank lines when only the content matters. Unchanged blocks collapse into compact context lines, so in long documents you only see the places where something actually happened.

For code there is syntax highlighting for JavaScript, TypeScript, Python, HTML, CSS, JSON, SQL and XML. Twelve ready-made examples - from a rental contract and a tax return to a ZUGFeRD invoice and an Nginx config - show you how the tool works instantly, with no need to paste your own text first.

The finished comparison exports as a unified diff, the format developers know and version control understands. Best of all, everything runs entirely in your browser. Not a single line of text is sent to a server, no contract, no code, no trade secret leaves your device. It is fast, works offline, and is exactly as confidential as a text comparison should be.