Base64 Encoder and Decoder

The tool converts text in one direction at a time, and the result updates as you type.

1. Paste or type your text into the Input box, or press Paste to take it from the clipboard.
2. Set Mode to Encode to turn text into Base64, or to Decode to read a Base64 string.
3. Tick URL-safe alphabet if the result will sit in a link, a filename or a JSON Web Token. It swaps + and / for - and _ and drops the = padding. The box only changes encoding.
4. Read the result in the output box. If a decode input is not valid Base64, a short error note appears instead.
5. Press Copy or Download .txt, or Use as input to send the result back in and flip the mode to check a round trip.

Decoding is forgiving: it accepts both alphabets, restores missing padding and ignores spaces and line breaks. Encoding produces one unbroken line, without the 76-character wrapping that email attachments use.

Base64 writes binary data using 64 printable characters: A to Z, a to z, 0 to 9, + and /. The encoder takes three bytes (24 bits), cuts them into four groups of six bits, and looks each group up in that alphabet. In Hel the bytes 01001000 01100101 01101100 become the groups 18, 6, 21 and 44, which are S, G, V and s. The leftover lo gives bG8, and a single = pads the final block to four characters.

One leftover byte earns two = signs, two leftover bytes earn one. Output length is the byte count divided by three, rounded up, then multiplied by four, so Base64 is about a third larger than the original. Text is converted to UTF-8 bytes first, so accents and emoji survive the trip.

Base64 exists to carry bytes through channels built for text. You meet it in HTTP Basic authentication, where user:pass becomes Authorization: Basic dXNlcjpwYXNz, in email attachments, and in the first two parts of a JSON Web Token. Decoding eyJhbGciOiJIUzI1NiJ9 here, which is URL-safe and unpadded, gives {"alg":"HS256"}. To read a whole token with its claims laid out, use the JWT decoder.

The standard alphabet contains +, / and =, which have jobs of their own inside URLs. The URL-safe form swaps them out: the text subjects?>> encodes as c3ViamVjdHM/Pj4= in the standard alphabet and c3ViamVjdHM_Pj4 in the URL-safe one. If you only need to escape characters for a query string, the URL encoder is the right tool. This page takes text only, so for pictures use the image to Base64 converter.

• Treating it as encryption. Base64 has no key. Anyone can reverse it, so it hides nothing. For a one-way fingerprint use the hash generator.
• A stray newline. A trailing line break is part of the input: Hello plus a newline encodes as SGVsbG8K, not SGVsbG8=. This is why command-line results sometimes disagree with a web tool.
• Expecting text from binary. If the decoded output shows diamond replacement characters, the data was probably an image, archive or other binary file, not text.
• Damaged input. A string whose length leaves a remainder of one after dividing by four cannot be valid, usually because characters were lost when it was copied.
• Two alphabets in one string. Mixing + with -, or / with _, is a sign that two sources were joined. The decoder accepts either, but other software may not.
• Assuming a file can be pasted. Base64 of a large file is a very long string. The converter handles plain text, so keep binary files for a tool that reads files directly.

A quick way to test a value is a round trip: encode it, press Use as input, switch Mode to Decode, and confirm the original text returns. If it does, the Base64 you are about to send is sound.