Question 1

How is Base62 different from Base64?

Accepted Answer

Both pack bytes into printable characters, but Base64 uses 64 symbols including the URL-unfriendly '+', '/' and the '=' padding sign, so a Base64 string usually has to be percent-encoded before it travels in a URL path. Base62 drops exactly those three characters and keeps only 0-9A-Za-z, which every URL parser leaves untouched. The trade-off is size: Base64 is a clean 6-bit packer, while Base62 needs BigInt division, so Base62 output is a few percent longer. For short links the URL-safety wins.

Question 2

Why do short-link services like bit.ly use Base62?

Accepted Answer

A short link maps a tiny public key to a long destination stored in a database. The database key is usually a sequential integer, and the shortest URL-safe way to print that integer is Base62, because all 62 characters can sit in a URL path with no escaping. A 7-character Base62 slug covers 62^7, about 3.5 trillion links, which is why you see codes like /3dPq9Zx. Sequential IDs encode to short strings early on and grow only one character at a time.

Question 3

Show me a concrete number example.

Accepted Answer

In Base62 the value 0 encodes to '0', 61 encodes to 'z' (the last single digit), and 62 rolls over to '10' exactly like 9 to 10 rolls over in base 10. So 125 encodes to '21' because 125 = 2 times 62 + 1, and the digit 2 then 1 gives '21'. Decoding reverses it: '21' is 2 times 62 plus 1, which is 125. Large IDs work the same way through BigInt, so a value like 9876543210 has a stable short form that round-trips exactly.

Question 4

How does text mode encode a string?

Accepted Answer

Text mode takes the UTF-8 bytes of your input, lines them up as one big-endian integer, and Base62-encodes that single number. Decoding does the reverse and rebuilds the original bytes. Because it works on raw bytes, it handles any content: 'Hello' encodes the same on every machine, and Chinese like 你好 or an emoji round-trips byte-for-byte. Note this is an encoding, not encryption, anyone can decode it, so it is for compactness and URL-safety, never for hiding secrets.

Question 5

What about Base58, the one Bitcoin uses?

Accepted Answer

Base58 removes four characters that humans misread: 0, O, I and l. That makes a copied-by-hand address less error-prone, which is why wallets and IPFS use it, and it preserves leading zero bytes as leading '1'. Base62 keeps those four characters, so it is denser and slightly shorter for the same value, but a touch easier to mistype. Pick Base58 when a person reads the string aloud or types it; pick Base62 when a machine generates and consumes it, as with auto-created short links.

Question 6

Is there a length limit, and what happens with invalid input?

Accepted Answer

Number mode accepts any non-negative integer of any size thanks to BigInt, so a 128-bit UUID-sized value is fine. Decoding rejects any character outside 0-9A-Za-z and tells you which character was bad, rather than silently producing garbage. Empty input simply yields an empty result instead of an error. Text mode round-trips arbitrary UTF-8; the only practical ceiling is the browser tab's memory, which you will not reach with normal link or token sized inputs.

Base62 Encoder and Decoder for Short URLs

What this tool does

Tool details

How to use

1. Input

2. Process

3. Copy / Download

How Base62 Encoder & Decoder fits into your work

Encoding jobs

Encoding checks

Good next steps

FAQ

Tool combos

Base62 Encoder and Decoder for Short URLs

What this tool does

Tool details

How to use

1. Input

2. Process

3. Copy / Download

How Base62 Encoder & Decoder fits into your work

Encoding jobs

Encoding checks

Good next steps

FAQ

Unit Converter

Percentage Calculator

Base32 / Base58 Encoder & Decoder

Base64 Encoder & Decoder

Number Base Converter

Text to Hex Converter

AES Text Encryptor

Affine Cipher Encoder & Decoder

Atbash Cipher

Bacon Cipher Encoder & Decoder

Base64 Block Deduplicator

Base64 Block Extractor