Codec lesson · Base32

Base32: represent bytes with a 32-symbol alphabet

Base32 uses uppercase A–Z and digits 2–7. Its alphabet avoids visually ambiguous 0, 1, 8, and 9, but the encoded result is longer than Base64.

Standard · RFC 4648 Round trip: Exact for canonical padded Base32 whose bytes are valid UTF-8 text.

Learning objective

Recognize RFC 4648 Base32, explain its padding, and validate canonical Base32 before decoding it as UTF-8 text.

Encode Keyboard groups UTF-8 bits into five-bit values, maps them to the RFC 4648 Base32 alphabet, and pads the result to an eight-character boundary with =.

How Encode Keyboard handles it

Convert text to UTF-8

All characters, including Vietnamese and emoji, begin as bytes.

Read five bits at a time

Each value from 0 to 31 selects A–Z or 2–7.

Require canonical blocks

Decode accepts uppercase symbols and trailing padding only when re-encoding produces the same value and the bytes form valid UTF-8.

Worked example

InputHi

OutputJBUQ====

Two input bytes leave a partial five-bit group, so four = characters complete the eight-character block.

Good uses

Learning binary-to-text encodings
Inspecting values explicitly documented as RFC 4648 Base32
Sharing non-sensitive text in a case-stable alphabet

Validation and common errors

Use uppercase A–Z and digits 2–7 only
Keep = only at the end and use the canonical count
The total canonical length must be a multiple of eight

Practice in your browser

This uses the same documented rules as the app and runs only in this tab.