Files, Binary & Compression

File checksumCheck each download against its reference

Add files, choose the algorithm named by the publisher, and enter an optional expected digest for each file.

File checksum

Files and expected digests

Drop files here, or choose several files.

Computed digests

File contents are read locally in your browser; they are not uploaded.

Up to 20 files, 2 GiB each. Read in 1 MiB chunks. MD5 and SHA-1 are for legacy compatibility.

Add a file to begin. Empty files are valid.

The list contains file names and computed hashes, including mismatches. It does not contain comparison verdicts or file contents.

File checksum helps you check downloaded installers, archives and disk images against separately published digests. Select the exact release and algorithm: a checksum for a different operating system or build will not match.

This file hash calculator reads the original bytes without converting line endings or unpacking archives. Use SHA256 file hash checks for references labeled SHA-256 and MD5 file hash checks only when an older release supplies MD5.

Why use File checksum?

Read large files in bounded chunks

File checksum feeds consecutive 1 MiB pieces into one running hash state. A 2 GiB image does not need a matching 2 GiB memory buffer, and the worker can be stopped without locking the page.

A large file passes through consecutive small blocks into one digest, with a progress indicator.

Bind each reference to the right file

An installer and its companion archive usually have different published values. Each input row keeps its own expected digest, so matching one file never marks every file as verified.

Two files feed separate digest comparisons; one comparison matches and the other differs.

Avoid the common comparison traps

01

Use the digest for the exact release

Select the algorithm before copying a reference into its file row. SHA-256 needs 64 hex characters; a checksum list line with a file name is not an accepted expected value.

02

Start a deliberate batch

Adding files appends them. Ctrl/Cmd+Enter runs the selected algorithm, and Cancel discards the unfinished batch. Changing even an expected value makes previous results stale; run again before exporting.

03

Keep the list next to the files

The downloaded GNU list uses relative ./file names and binary-mode markers. Place it beside the original files and use the matching checksum program. Example and Clear first show a preview; Undo restores their previous inputs and results.

Inspect bytes, references and batches

Four exact algorithms

Choose MD5, SHA-1, SHA-256 or SHA-512 for one entire batch. Their hex lengths are 32, 40, 64 and 128 characters. The tool never guesses the algorithm from a pasted value or silently substitutes one of equal length.

An independent verdict per file

Leave an expected value blank to calculate only. Otherwise the whole hex digest is compared after removing surrounding whitespace and ignoring letter case. A well-formed but different value produces a mismatch, while malformed input is an error.

Progress across all selected bytes

Files run sequentially in a worker with byte-based progress. Empty files have valid digests, and an all-empty batch completes without dividing by zero. Read failures stop the batch; incomplete results cannot masquerade as a finished list.

A reproducible GNU list

Export includes the selected algorithm’s computed digest and each relative file name, even if a supplied reference did not match. Backslashes and line breaks are escaped, and duplicate names are rejected before reading to prevent ambiguous entries.

Limits and what a match can tell you

Integrity is relative to the reference
A matching digest does not identify the publisher, verify a signature or scan for malware. Obtain the reference through a trustworthy channel. MD5 and SHA-1 remain available for older checksum lists, not modern security decisions.
A finite browser workload
A batch holds at most 20 files and each file may be at most 2 GiB. Larger selections are rejected before reading. The fixed 1 MiB chunk size bounds the input buffer; CPU time still grows with file size and depends on the device.
Files retain their exact bytes
No text decoding, Unicode normalization, BOM removal, newline conversion or ZIP extraction occurs. There is no salt, key, HMAC, password hashing or Base64 output on this page. A ZIP checksum belongs to the archive itself.
Names must identify a single file
The list does not preserve directory trees. Same-name files from different folders must be processed in separate batches or renamed first. Slash, NUL and dot-only path names are rejected; backslash, CR and LF use GNU escapes in export.
No remote input or resume
Directory recursion, URL fetching and checksum-list import are not supported. References are entered individually as complete hex. Cancel stops current work; the next run starts from the beginning, and refreshing the page discards the selection.

File checksum questions

Which digest algorithms can I use?

The selector offers MD5, SHA-1, SHA-256 and SHA-512, with SHA-256 selected initially. One algorithm applies to all files in a batch. Choose the exact algorithm specified alongside the reference; similar-looking strings are not interchangeable.

Are selected files sent to a server?

No. Selection gives the browser access to those files for this calculation. The generated checksum list includes their names and hashes, so review the file names before sharing that list.

Why do normalized and exact checksum comparisons disagree?

A comparison that treats letter case or surrounding spaces as data can disagree with one comparing hex bytes. This page trims only the outside whitespace and ignores hex letter case. It rejects internal spaces, prefixes and pasted checksum-list lines.

How should I investigate a checksum mismatch?

Check the release, platform, file name and chosen algorithm first. Recopy the full published value and, if necessary, download the file again. An HTML error page saved with an archive extension is still a different set of bytes; the tool does not inspect its format.

Does a matching digest establish that a download is safe?

No. Matching a digest supplied by the same untrusted source proves no independent trust. This tool checks byte consistency against your reference and does not authenticate a publisher, validate a signature or detect malware.

Still have a question?

Check your next download

Use File checksum with the built-in example to see one match and one mismatch before checking your own download.