Hash text or a file with SHA-256, SHA-384 or SHA-512, or sign it with HMAC. Output as hex, Base64 or Base64url. Everything runs on your own device.
Computing…
Large inputs take a moment.
—
—
A hash is not encryption. It cannot be reversed, and it also cannot protect anything — it is a fingerprint, not a lock.
A matching hash is not proof of authenticity. Anyone who can change a file can change the digest published beside it. That is what HMAC and signatures are for.
Hash Generator runs entirely on your device — nothing is uploaded.
You’re on the Free plan. Compare plans
A hash is a fixed-length fingerprint of some input. The same input always produces the same digest, a one-character change produces an entirely different one, and there is no practical way to work backwards from the digest to the input. That makes hashes useful for checking that a download arrived intact, that two files are identical, or that a value has not been altered in transit.
Hashing is not encryption. Encryption is reversible by design — that is the point of a key — whereas a hash throws information away and cannot be undone. Anyone who tells you a password is "encrypted with SHA-256" has described something that does not exist. A digest on its own also proves nothing about who produced it: an attacker who can change a file can change the published digest beside it just as easily. That is what HMAC is for, and why this tool offers it.
HMAC adds a secret key to the construction, so only someone holding the key can produce a matching tag. It is what webhook signatures from Stripe, GitHub and most other providers use, and verifying one locally is the fastest way to find out whether your own implementation agrees. The key stays in the page: it is never logged, stored or sent anywhere.
SHA-256, SHA-384 and SHA-512 are the recommended choices here. SHA-1 is included because real systems still emit it — Git object IDs, older certificates, legacy APIs — and you sometimes need to reproduce one to compare. It is marked as unsuitable for anything new, because practical collisions against it have been demonstrated since 2017. MD5 is not offered at all; it has been broken for long enough that there is no honest reason to generate a new one.
The hashing uses the Web Crypto API built into your browser, which is the same implementation your own code would call. Files are read with FileReader and hashed in memory, so even a large archive never leaves the machine.
No, and the difference matters. Encryption is reversible with a key; hashing destroys information and cannot be reversed at all. A hash is a fingerprint, not a locked box. If you need to get the original value back, you need encryption, not hashing.
Only that it matches the digest you compared against. If an attacker controls both the file and the published digest, both will match and nothing is proven. Authenticity needs a key, which means HMAC or a digital signature — a bare hash cannot do it.
Because collisions in MD5 are trivial to produce and have been for years, and offering it would invite someone to use it for something that matters. If you need to reproduce an existing MD5 value, use a dedicated checksum utility and treat it as a legacy identifier rather than a security property.
Collisions against SHA-1 were demonstrated in practice in 2017, so it cannot be relied on to prove that two inputs are different. It is here because systems in the wild still produce it and you sometimes have to match one. Never choose it for new work.
No. Passwords need a deliberately slow algorithm built for the purpose — bcrypt, scrypt or Argon2 — with a per-user salt. A single pass of SHA-256 over a password can be brute-forced at billions of guesses per second on ordinary hardware.
No. Hashing happens in your browser through the Web Crypto API, files are read locally, and neither the input, the file contents nor the key is uploaded, logged or recorded in analytics.
Almost always a difference in what is being hashed rather than how. A trailing newline, CRLF against LF line endings, or a different text encoding all change the bytes and therefore the digest. Hashes are computed over bytes, not over what the text looks like.