Base64 to Image Converter

This is the reverse of the Image to Base64 tool, and it turns up in more situations than you would expect: debugging an API response, checking whether an image survived a database round trip, recovering an asset from an HTML email, or verifying that a build step inlined the right file.

What the decoder accepts

Real-world Base64 rarely arrives clean. The parser handles:

• a complete data URI with any MIME type;
• a CSS url("data:image/png;base64,…") wrapper, quoted or unquoted;
• bare Base64 with no prefix at all;
• strings broken across lines, which happens constantly when copying from terminals, logs and email;
• missing trailing = padding, which some encoders omit.

Whitespace is stripped and padding is restored before decoding, so a payload that would fail a strict atob() call still works here.

Reading the failure messages

Invalid character means something outside the Base64 alphabet is present. If you see % characters the string is still URL-encoded — run it through the URL decoder first. If you see - or _ it is base64url, which this tool translates automatically.

Truncated data means the payload decoded but the image structure is incomplete — typically a copy that stopped early. Compare the character count against what the source claims.

Unrecognised format means the bytes decoded cleanly but do not begin with a known image signature. The data may be a document, an archive or simply not Base64 image data.

Why magic-byte detection matters

A data URI's declared MIME type is written by whoever created it, and it is frequently wrong — especially after a file has been renamed or passed through a system that guesses types from extensions. Reading the first few decoded bytes tells you what the image actually is, which is why the detected format is shown separately from the declared one and used to pick the default download extension.