Base64 vs Base64URL: Encoding Unicode Text Correctly

Base64 is often described as turning text into “encrypted-looking” characters. That is misleading. Base64 is a predictable way of representing bytes using a limited alphabet. Anyone can decode it, with no password.

What happens when you encode a string?

First convert the string to a byte sequence. On modern websites, UTF-8 is usually the right encoding. Second, divide those bytes into 6-bit groups and map them onto 64 printable symbols. The result works with systems that prefer ASCII-only data.

Try Hello in the Base64 tool. Its UTF-8 bytes become SGVsbG8=. Decode that output and you get the original text.

Why Unicode is a common pitfall

The classic browser functions btoa() and atob() operate on byte-like strings, not arbitrary Unicode text. Feeding a Chinese or emoji character directly to btoa can fail. Vyblet first encodes to UTF-8 bytes, allowing inputs such as 你好 🌏 to round-trip correctly.

Standard Base64 and Base64URL

Standard Base64 can contain +, / and final = padding. Base64URL substitutes - and _ and often omits padding. It is frequently used inside tokens and other web-safe formats. Neither is inherently encrypted or secret.

When Base64 helps—and when it doesn't

1. Helpful: moving small binary payloads through text-based transports and inspecting examples in developer APIs.
2. Less helpful: compressing data. Base64 normally increases size by about one third.
3. Not suitable: protecting passwords, tokens or sensitive records. Use a well-designed encryption system for confidentiality.

Debugging decoding errors

Invalid characters, the wrong variant, truncation or non-UTF-8 source bytes commonly cause errors. If you're decoding a PNG, ZIP or other binary content, a text decoder is the wrong final step: that content is bytes, not readable text.

FAQ