Base62 Encoder / Decoder

How to use

1. 1Pick Encode or Decode at the top, or use Swap to flip the current direction.
2. 2Choose the alphabet that matches your system: GMP / standard (0-9 A-Z a-z) is the most common, with inverted and truncated orders available for sources that differ.
3. 3When encoding, choose UTF-8 text or Hex bytes for the input. Hex lets you encode arbitrary binary such as a 16-byte UUID or a hash without character-set issues.
4. 4When decoding, choose UTF-8 text or Hex bytes for the output. The tool falls back to hex automatically when the decoded bytes are not valid UTF-8.
5. 5Load a quick-conversion preset (Hello, World!, a UUID to shorten, a hex value with a leading zero byte) to see a worked example and confirm the alphabet.
6. 6Click Copy Base62 (or Copy output when decoding) to grab the result. Click Clear to reset the input.

About this tool

Base62 Encoder / Decoder converts between raw bytes and Base62 entirely in your browser. Base62 uses the 62 alphanumeric characters (the ten digits, twenty-six uppercase letters, and twenty-six lowercase letters) and no symbols at all, which is exactly what makes it popular: the output is safe to drop into a URL path, a query string, a filename, or a CSS identifier without percent-encoding and without the +, /, and = characters that standard Base64 introduces. It is the encoding behind a great many URL shorteners (the short code after the slash), compact opaque IDs, hashed asset filenames, and shrunk representations of large counters or hashes. This tool uses the byte-oriented big-number convention, the same one the widely used base-x and base62 libraries implement: the input bytes are read as a single base-256 integer and re-expressed in base 62 by repeated long division, so inputs of any length encode correctly without relying on BigInt. That is a different operation from Base64, which slices the input into fixed 6-bit groups; Base62 is not a drop-in replacement for Base64 and the two produce different strings, which this tool makes explicit. Encoding accepts either UTF-8 text or a hex byte sequence, so you can encode a string like Hello, World! or a 16-byte UUID without character-set ambiguity, and decoding renders the result as UTF-8 text with an automatic, clearly labelled hex fallback when the bytes are not valid UTF-8. Because real systems disagree about which character order counts as Base62, three alphabets are offered and shown in full: GMP / standard (digits, then uppercase, then lowercase), which most short-ID systems and the base62 npm default use; inverted (digits, then lowercase, then uppercase); and truncated (uppercase, then lowercase, then digits), a Base64-style order with the symbols removed. Decoding with the wrong alphabet either errors on an unexpected character, with the exact position reported, or silently yields different bytes, so the selector makes the choice deliberate. The leading-zero rule is handled in both directions: each leading zero byte becomes one leading 0 character and the byte count is restored exactly on the round trip, so a hex value such as 00ff does not lose its leading byte. The decoder ignores whitespace in pasted input so a wrapped or multi-line string works without cleanup. Useful for shortening UUIDs and database IDs into compact URL-safe codes, decoding a short link back to its underlying bytes, generating readable identifiers, comparing Base62 against Base64 and Base58, and any developer task that needs symbol-free alphanumeric encoding. Everything runs locally on your device using TextEncoder, TextDecoder, and plain integer math. Nothing you paste here is uploaded, logged, or shared.

Free to use. Works in your browser. No signup, no login.