Eleven small tools I reach for when something looks off at 2 AM. Decode the token the service handed back. Check the hash before you trust the download. Scan the paste before it goes into the repo. Nothing here leaves your browser.
I use this when a secret or a token comes back as a blob and I need to see what is inside it. Handles full Unicode text, not just ASCII.
Paste a token and read its header and payload as plain JSON. I do this when an API call comes back 401 and I need to know what the token actually claims. This decodes only. It checks no signature.
A decoded token is not a trusted token. Anyone can read these claims. The signature is what proves who issued it, and this tool does not verify it.
Random version 4 UUIDs, built from your browser's secure random numbers. I use them for correlation IDs, test fixtures, and anywhere a throwaway unique key is needed.
Random passwords from your browser's secure random numbers. Pick a length and the character sets you need. Nothing is saved or sent anywhere, so copy it into your password manager right away.
Hash text with SHA-1, SHA-256, SHA-384, SHA-512, or MD5. I hash a string when I need a quick fingerprint of it, for example to compare a config value across two clusters. MD5 is computed here in plain JavaScript because browsers do not ship it. Do not use MD5 or SHA-1 for anything that must resist tampering.
Paste a PEM certificate and read what it actually says: subject, issuer, dates, key type, and names it covers. I do this before a deploy when a cert is about to expire or a handshake fails and nobody remembers what was issued.
Paste a PEM certificate signing request and check what was asked for before it goes to the CA. Subject, key type, and requested extensions, so a typo in the common name gets caught here and not after issuance.
Paste text and the hash you were given, and see if they match. I use this to check a download or a pasted secret against its published checksum before I trust it.
Paste code, config, or a log snippet and this scans it for leaked secrets: cloud keys, tokens, private keys, and passwords in plain text. It runs fully in your browser. Paste it here and it never leaves this page, which is the whole point.
This catches common patterns, not every secret. A real credential can hide in a shape this list does not know. Treat a clean result as "nothing common found", not "safe to commit". If anything is flagged, revoke it and rotate it. Redacting a file after the fact does not un-leak a key.
Paste a public key line from authorized_keys or an .pub file and see what it is: key type, size, comment, and fingerprints. I use this when a login fails and I need to know whether the key on the server is even the key I think it is.
Type a password and see how much guessing work it really costs an attacker. Entropy in bits, the character sets it uses, and a rough crack time. Type it here only if you are comfortable doing so. Nothing is stored or sent, and the field clears with the rest of the page.
Entropy here measures the size of the guessing space if the password were random. A real word with a digit on the end scores far worse than its characters suggest, because attackers try words first. Length beats cleverness.