Base32 encoder / decoder

What Base32 is and why it exists

Base32 is an encoding scheme defined in RFC 4648 that represents any binary data using only 32 characters: the letters A-Z and the digits 2-7, plus = as padding. Each character carries 5 bits, so every 5 input bytes become exactly 8 output characters. Why another scheme when Base64 already exists? Because Base32 has two superpowers Base64 lacks: it is case-insensitive (no two characters differ only in case) and it avoids ambiguous characters — the alphabet contains neither 0 nor 1, so you can never confuse them with O or l. That makes it ideal for anything a human has to read, dictate over the phone or type by hand: the TOTP/2FA secret keys in Google Authenticator, names encoded in DNS (which is case-insensitive by design), and ULID identifiers, which use Crockford's Base32 variant.

How it works: 5 bytes become 8 characters

The algorithm takes the input bytes (here, your text encoded to UTF-8 with TextEncoder, so accents, ñ and emoji come out perfectly) and reads them as one continuous stream of bits. Each group of 5 bits is a number between 0 and 31, which maps to one character of the alphabet. Since 5 bytes = 40 bits = exactly 8 groups, the natural Base32 block is 8 characters. When the input is not a multiple of 5 bytes, the last group is completed with zero bits and the output is padded with = up to 8. The classic RFC example: foobar encodes to MZXW6YTBOI======, and fo to MZXQ====. You can verify it right now in the tool above.

Base32 vs Base64 vs Hex at a glance

The rule of thumb: if only a machine will ever read the result, use Base64 (more compact) — that is what our Base64 encoder and Base64 decoder are for. If a human will read, dictate or type it, or the channel ignores case (like DNS), Base32 is the right choice.

Where you run into Base32 in real life

• TOTP/2FA keys. The secret you scan as a QR code into Google Authenticator, Authy or 1Password is Base32. If the QR fails, the key you type by hand (like JBSW Y3DP EHPK 3PXP) is too.
• DNS and DNSSEC. Domain names ignore case, so encoding hashes into records (NSEC3) requires a case-insensitive alphabet.
• ULID. Sortable unique identifiers that use Crockford's Base32 (a variant with a different alphabet) to stay compact and readable.
• Human-friendly tokens. Activation codes, license keys and coupons that a person has to type without frustration.

Reference: the Base32 alphabet (RFC 4648)

The = padding carries no value: it only completes the 8-character block when the input is not a multiple of 5 bytes.