What this tool calculates
Choose the algorithm named by the publisher, then select or drop the actual local file. The default is MD5; the same queue also supports SHA-1, SHA-256, SHA-384, SHA-512 and CRC32. Files are processed sequentially in 2 MiB chunks. This avoids loading several complete large files at once, but processing still uses your device's CPU and memory.
Each file has its own name, byte size, progress and result. A zero-byte file has a valid digest. Cancel stops active work; queued files can be cancelled or removed independently. Retry uses the same selected local file. Changing the algorithm invalidates old results and recalculates eligible files; cancelled entries wait for Retry.
Paste only the expected hexadecimal value into comparison, not a filename or a whole command-output line. A malformed value is different from a valid mismatch. Obtain that value from a trusted source. Matching checksums detect many accidental changes but do not certify that software is safe or that its publisher is authentic.
Worked example
UTF-8 · abc · MD5
900150983cd24fb0d6963f7d28e17f72
How to calculate and compare
- Select Algorithm before adding files. MD5 is the default on this page; choose the algorithm named by the publisher's reference.
- Click Choose files or drag local files into the drop area. Processing starts automatically, one file at a time. Example adds a three-byte
example.txtcontainingabc, with no final newline. - Each file shows its name, byte size, progress and state. Cancel a running or queued item, remove it, or retry a cancelled calculation. Changing the algorithm clears old results and recalculates eligible files; cancelled items stay cancelled until Retry.
- Paste only the reference's hexadecimal value into Compare expected value (optional) and click Check. Do not paste a filename, an algorithm prefix or a whole terminal-output line.
- Use Copy for one full digest, its Download control for that file's result, or Download TXT / CSV below the queue for all completed results.
An empty file is valid. Hashing the saved file is different from pasting its displayed text: encodings, metadata and line endings are part of the original bytes. TXT and CSV contain file names, algorithms and complete results, not file contents. A TXT or CSV export is a result table, not a checksum manifest accepted by md5sum --check.
Limits, privacy and next steps
The queue accepts up to 50 files, each no larger than 1 GiB. Files are read in sequential 2 MiB chunks in a Web Worker, not uploaded for hashing. Large files may take longer on phones; cancel and use a desktop utility if the device struggles. Removing or clearing an item releases the application's reference to that file. Downloads are local records that you manage yourself. Normal website requests still reach the host.