Base64 Decoder

How the Base64 Decoder works

The decoder reads the input one character at a time. Spaces and line breaks are ignored, both the standard (+ /) and URL-safe (- _) alphabets are accepted, and missing = padding is added back automatically. If the input starts with a data URI header such as data:image/png;base64,, the header is removed and its MIME type is remembered.

Any other character stops decoding with an error that gives its position (and line and column for multi-line input), so you can find a stray quote or a character damaged by copy and paste. Input that ends in the middle of a character group is reported as incomplete.

The decoded bytes are then checked. Valid UTF-8 text is shown in the output box. Bytes that are not valid UTF-8, contain null characters or start with the signature of a known file type (PNG, JPEG, GIF, WebP, PDF, ZIP or Gzip) are treated as a file: you get a Download decoded file button with a matching extension, and images are previewed.

How to use the Base64 Decoder

1. Paste Base64 text or a data URI into the input box, open a text file that contains Base64, or select Load example.
2. Select Decode Base64, or press Ctrl + Enter (Cmd + Enter on a Mac).
3. If the result is text, it appears in the Decoded text box, ready to copy or download.
4. If the result is binary data, select Download decoded file. Images are also previewed on the page.
5. If the input contains an invalid character, the error message gives its position and the cursor moves to it.

Example

This input:

SGVsbG8sIFdvcmxkISBDYWbDqSDimJU=

is decoded to the text:

Hello, World! Café ☕

The URL-safe form without padding, SGVsbG8sIFdvcmxkISBDYWbDqSDimJU, gives the same result. Input such as SGVsbG8h! is rejected with Invalid Base64 character '!' at position 9.

Common use cases

• Reading a Base64 value from an API response, a configuration file or a Kubernetes secret.
• Turning a data URI from HTML or CSS back into an image file.
• Saving a PDF or ZIP attachment that an API returned as a Base64 string.
• Decoding the parts of a JWT or other URL-safe Base64 strings that have no padding.
• Finding a stray character that makes another program reject a Base64 string.

Common errors and how to fix them

Invalid Base64 character

Something in the input is outside the Base64 alphabet. Common causes are a line break or space introduced by copy and paste, or a URL-safe - or _ while the standard alphabet is selected. Remove the stray characters, or switch alphabets.

The Base64 data is incomplete

Base64 is read in groups of four characters and the input does not end on a complete group. The value was almost certainly truncated when it was copied; copy it again in full, including any trailing = padding.

Too much padding, or the input contains only padding characters

There are more = characters than a valid value can end with, which means the data is corrupted or two values have been concatenated. Re-copy the original.

This data URI is not Base64-encoded

The data URI header has no ;base64 marker, so its payload is percent-encoded rather than Base64. Use the URL Decoder for that data instead.

Frequently asked questions