Capability overview
What does reports and trusted preprocessors do in Diff Anything?
Pro exports self-contained HTML or JSON reports, unified diffs for text, and changed-row CSV for CSV comparisons. It can also run an explicitly confirmed local executable directly—without a shell—on two text sources before comparison.
Diff Anything chooses a comparison engine that understands the inputs. Code and text receive language-aware highlighting and moved-block detection, text adds an interactive three-way merge workspace, JSON and other structured formats compare semantic paths, folders can synchronize a reviewed plan with rollback and undo, PDFs connect rendered pages with extracted text, and images offer overlay, swipe, blink, heatmap, and changed-pixel views.
In the workflow
Save or automate
Use Pro to export a report or retain a profile, invoke the CLI, or wire the comparison into Git and CI.
Diff Anything keeps the source pair, selected mode, comparison rules, result, and review state together so the same decision can be revisited or automated.
Useful in practice
Where this capability earns its place.
Related manual
Profiles, rules, review state, and reports
Turn a useful one-off comparison into a repeatable review with explicit ignore rules, notes, and exportable evidence.
The guide covers save a profile, use ignore rules carefully, review semantic differences, apply supported changes, export a report, use live watch deliberately. It is the best next step when you want to move from the product overview into a repeatable setup.
Read the guideDirect answers
Questions about reports and trusted preprocessors.
What does reports and trusted preprocessors do in Diff Anything?
Pro exports self-contained HTML or JSON reports, unified diffs for text, and changed-row CSV for CSV comparisons. It can also run an explicitly confirmed local executable directly—without a shell—on two text sources before comparison.
Do compared files leave the computer?
No. Comparison is local. The application has no account, cloud comparison service, analytics, or telemetry.
How does automatic mode choose a comparison?
Matching structured formats use their semantic engine, two folders use folder comparison, and other compatible inputs fall back to text. You can always choose a concrete mode.