Base64 Decode
Decode standard padded Base64 back to readable UTF-8 text, including emoji and Indian scripts. Strict checks catch bad padding and broken input. Free tool.
- 100% free
- No sign-up
- Runs in your browser
- Updated
Your result
Your converted output and its statistics appear here.
How this was calculated
About Base64 Decode
Found a string like SGVsbG8= in an email header, API response or configuration file? This Base64 Decode tool turns standard Base64 back into readable text. It decodes the bytes and interprets them as UTF-8, so emoji, accented letters and scripts such as Devanagari come back exactly as they were written.
How to use Base64 Decode
- Paste the Base64 string into the Standard padded Base64 box.
- Or press Open a text file to load Base64 stored in a local .txt, .csv, .json, .md, .html, .xml or .log file of up to 5 MB.
- Press Convert to decode it.
- Read the decoded text in the result area; if an error appears, check the padding or alphabet.
- Press Copy result to use the text, Download to save it as decoded.txt, or Reset to clear the box.
Key features
- Decodes standard Base64 (A–Z, a–z, 0–9, + and /) with required = padding.
- Interprets the decoded bytes as UTF-8, supporting emoji and all Unicode scripts.
- Ignores spaces, tabs and line breaks, so wrapped Base64 decodes as is.
- Rejects missing or incorrect padding and noncanonical padding bits with a clear message.
- Detects bytes that are not valid UTF-8 text instead of showing replacement symbols.
How it works
Validate alphabet, padding, and canonical padding bits; decode bytes; decode UTF-8 in fatal mode.
Base64 Decode examples
Short padded string
Input: SGVsbG8=
Result: Hello
Text with an emoji
Input: SGVsbG8g8J+MjQ==
Result: Hello 🌍
Important notes before you rely on the result
- Everything runs client-side, so it is safe to paste API responses, but avoid pasting live passwords or private keys into any website.
- Base64 is an encoding, not encryption: anyone can decode it.
- JSON is parsed, never executed. Comments and trailing commas are not valid JSON and are reported as errors.
Limits and supported formats
- Standard padded Base64. The decoded bytes must be valid UTF-8 text.
Privacy
Runs in your browser. Nothing you enter is uploaded or saved.
Base64 Decode: frequently asked questions
Why do I get a Use standard padded Base64 error?
The input contains characters outside the standard alphabet, or its length and padding are wrong. Common causes are missing = signs at the end, URL-safe characters - and _, or stray symbols copied with the string. For example, SGVsbG8 fails because it should end with =. Add the padding or replace - with + and _ with /, then try again.
What does noncanonical padding bits mean?
In valid Base64, the unused bits in the last character before the padding must be zero. If they are not, the string was probably edited or generated incorrectly, even though some lenient decoders would accept it. For example, SGVsbG9= is rejected while SGVsbG8= decodes to Hello. Check that the string was copied completely and has not been altered.
Can it decode URL-safe Base64 or JWT tokens?
Not directly. URL-safe Base64, used in JWTs, uses - and _ instead of + and / and often leaves out padding. Convert it first: replace - with +, _ with /, and add = signs until the length is a multiple of four. Each JWT section is decoded separately, and the signature part is binary, so it will not decode as text.
Why does it say the bytes are not valid UTF-8?
The Base64 itself is fine, but the decoded bytes are not readable text. This usually means the original data was binary, such as an image, a compressed file or an encrypted value, or text in an older encoding such as Windows-1252. This tool only returns UTF-8 text, so it reports an error rather than displaying broken characters.
Do spaces and line breaks in the input matter?
No. ASCII spaces, tabs and line breaks are removed before decoding, wherever they appear in the string. That means Base64 wrapped at 76 characters, as in email messages, or split across several lines decodes without any cleanup. Other characters, including non-breaking spaces or quotation marks copied with the string, are not ignored and cause an error.