Base64 Encoder / Decoder
Encode Unicode text as Base64, or decode standard and URL-safe Base64. Everything is processed in this browser; decoded binary can be downloaded without treating it as text.
Convert text or encoded data
Maximum: 2 MiB of UTF-8 or decoded data. Base64 changes representation; it does not encrypt or protect secrets.
How to use Base64 Encoder / Decoder
- Paste ordinary text and select Encode, or paste encoded data and select Decode.
- For web tokens, choose URL-safe encoding and optionally omit padding. Decoding accepts either alphabet and validates remaining padding bits.
- Copy or download UTF-8 output. If decoded bytes are not valid UTF-8, use the binary download instead.
Method and limits
Text is encoded as UTF-8 before Base64 conversion. ASCII whitespace in encoded input is ignored. Invalid characters, misplaced padding and nonzero unused bits are rejected. A data: URL prefix is not part of the accepted input. Decoded text downloads and unedited output-to-input round trips preserve original line endings; the text editor displays LF line breaks, and editing uses that displayed text. No input is sent to a conversion service or stored.
Reference: RFC 4648 Base64 specification
Frequently asked questions
Is Base64 encryption?
No. Anyone can decode it. Use encryption or an appropriate secure transport when you need confidentiality.
Does it support emoji and non-English text?
Yes. Text conversion uses UTF-8, including emoji and non-Latin characters.
Why are decoded bytes not shown as text?
Base64 can represent images and other binary files. Invalid UTF-8 is kept as bytes instead of silently replacing data with unreadable characters.
Worked examples: Unicode text and binary output
Paste café without quotation marks, leave the encoding options unchecked and select Encode to Base64. The result is Y2Fmw6k=. Use output as input and decode: the text is café again. The accented character is represented by UTF-8 bytes before encoding; treating each visible character as one byte would give the wrong result.
Now decode /w==. It represents one byte with hexadecimal value ff, which is not a complete valid UTF-8 text sequence. The tool offers a decoded-byte download instead of inventing a replacement character. A binary result is not necessarily an error; Base64 can represent arbitrary file contents. This tool can decode pasted binary data but its encoder accepts text, not file uploads.
Which options should I select?
Standard Base64 uses + and /. URL-safe Base64 replaces those with - and _. Choose the alphabet required by the receiving application. Only omit = padding when that format permits it; adding or removing characters at random is not a reliable fix for an invalid token.
Fix a decoding error without losing the original
- Paste only the encoded payload. Remove a wrapper such as
data:image/png;base64,before decoding. - Check whether a copied plus sign became a space in a query string. Whitespace is ignored here, so a lost plus sign cannot be reconstructed automatically.
- A single Base64 character cannot encode a complete byte. Truncated input and nonzero unused padding bits are rejected.
- For a dot-separated token, decode the appropriate segment separately; this is not a token-signature verifier.
More reproducible samples, line-ending checks and exact option differences are in the UTF-8 and binary Base64 guide. Encoding makes data portable, not confidential.
Examples and practical checks
Understand the result before using it elsewhere.
- Decode Base64 without confusing text and binary bytesWorked example · 5 min read
- Base64 is not encryption: what it does and when to use itWorked example · 5 min read
- Explore Developer ToolsCompare tools and their limits
- How we test these toolsMethods, verification and limitations