Duplicate File Finder
A matching filename or hash alone does not prove that two files are interchangeable. Select local files or a folder and this tool groups candidates by size, calculates SHA-256 in bounded chunks, then compares candidate bytes before declaring an exact match. Review each confirmed group and choose a representative and any additional files to keep. No source file is moved or deleted.
Key features
- Group files by actual content, even when names or folder paths differ
- Use file size and SHA-256 to narrow candidates, then compare bytes to confirm exact matches
- Keep same-name files with different content in separate groups
- Report unreadable or interrupted files as unresolved instead of claiming a match or mismatch
- Choose a representative and mark other copies to keep before exporting a review plan
- Read bounded chunks in the browser; leave the original files untouched
How to use
- Select multiple files or choose a folder in a browser that supports folder selection; alternatively load the sample set.
- Run the comparison and review confirmed duplicate groups and any unresolved files.
- Choose one representative in each confirmed group and mark any extra copies that should also be kept.
- Download the JSON review plan and inspect it before making any separate changes to the originals.
Use cases
- Find an unchanged document saved under different names in several folders
- Check whether two identically named exports actually contain the same bytes
- Prepare a keeper list for a backup cleanup without deleting any files here
- Identify exact copies independently of image appearance or file extension
Frequently asked questions
Does this tool delete duplicate files?
No. It only compares files and downloads a review plan with your keeper choices. It never renames, moves or deletes the originals. Make any file changes separately after reviewing the report.
Is a matching SHA-256 value enough to confirm a duplicate?
No. Size and SHA-256 narrow the candidate set; the tool then compares the actual bytes before listing files as an exact duplicate group. A file that cannot be fully read is reported as unresolved rather than silently assigned.
What if names match but contents differ, or names differ but contents match?
Grouping is based on content, not filenames. Different names with the same bytes can share a group; matching names with different bytes remain separate. Relative paths help you recognize each source in the review plan.
How is this different from the hash generator or similar photo finder?
The hash generator displays checksums for individual files. This tool clusters many files and confirms matching bytes before asking which copies to keep. The similar photo finder estimates visual resemblance; visually similar photos may have different bytes and are outside exact-duplicate matching.
What are the limits and folder requirements?
Select up to 500 files, at most 256 MiB per file and 1 GiB combined. Reads use 256 KiB chunks. Folder selection depends on browser support and your permission; selecting files directly is an alternative. The comparison cannot scan files you did not select.
Are my files or choices sent to the server?
No. Comparison runs in this tab. The downloaded JSON plan contains metadata and your choices, not file contents. It can reveal filenames and paths if you share it.
Privacy
Selected file bytes are read locally in this browser tab for comparison. The files are not uploaded, stored in a browser database or included in the exported review plan. The plan may contain the selected relative paths, sizes, digests and your choices, so inspect it before sharing.
Comments & questions