First decide whether your input represents text or arbitrary bytes. This site’s string converter uses UTF-8 and standard Base64; the file tool encodes uploaded bytes and does not decode Base64 back to a downloadable file.

Test input and settings

Hello, 世界!
Open the tool and follow the steps below

Reproduce a text roundtrip

Choose Encode and paste the sample without a trailing newline. Expect SGVsbG8sIOS4lueVjCE=. Choose Decode and paste that output; the result must be Hello, 世界! including its space and punctuation. A newline is another byte, so adding one changes the output.

Diagnose a decoding failure

/w== represents byte FF, which is not valid UTF-8 on its own. The text tool reports a failure even though that is valid Base64. Do not treat every text-decoding failure as corrupt Base64; use a byte-oriented decoder for binary payloads. This site’s file encoder cannot perform that reverse operation.

Match the surrounding protocol

A data: URL contains a prefix that is not part of the Base64 payload. Base64url uses - and _ instead of + and /; do not paste a JWT segment into this standard decoder and assume compatibility. Select the protocol’s decoder. Standard padded output length is 4 × ceil(byteLength / 3), so a one-byte input uses four characters, not a 33% increase.

Completion criteria and limits

Compare the recovered text exactly. Encoding does not encrypt credentials or authenticate a message; Basic authentication still needs a protected transport.

Technical reference

Back to tasks and guides · Report an incorrect instruction or result