What Base64 actually is
Base64 rewrites arbitrary bytes using only 64 safe characters (A–Z, a–z, 0–9, +, /, with = as padding). It exists so binary data can travel through channels that only handle text — HTTP headers, JSON payloads, email, config files. It is not encryption: anyone can decode it instantly.
The UTF-8 trap
Base64 encodes bytes, not characters. Naive browser code hands non-ASCII text straight to btoa() and throws an error, or worse, quietly mangles it. This tool converts your text to UTF-8 bytes first, so 안녕 and 👋 come back exactly as they went in. If you've ever decoded a string into mojibake, this is why.
URL-safe mode
Standard Base64 contains + and /, which mean something else inside a URL or filename, and = padding, which gets escaped. URL-safe Base64 swaps them for - and _ and drops the padding. You'll see it in JWTs, OAuth tokens and query parameters — decode those with URL-safe mode on.
Data URIs
A data URI embeds content directly in a link or stylesheet instead of pointing at a file: data:text/plain;base64,SGVsbG8=. Handy for tiny assets and quick tests, wasteful for anything large — Base64 makes data about 33% bigger.
FAQ
Is my text sent to a server?
No. Encoding and decoding happen locally in your browser with JavaScript. Nothing leaves the page, and there's no sign-up.
Is Base64 secure?
Not at all. It's an encoding, not encryption — reversing it takes one click. Never use it to hide secrets; use it to move bytes safely through text-only channels.
Why does decoding fail on my string?
Usually one of three things: it's URL-safe Base64 (turn the option on), it lost its padding when copied, or it wasn't valid Base64 to begin with. This tool tolerates missing padding and whitespace.
Where you'll actually meet Base64
Once you recognize it, Base64 is everywhere: the Authorization: Basic header packs a username and password as Base64, a data URI embeds an image or font as Base64 inside CSS, an email attachment is Base64 in transit, and the header and payload of a JWT are Base64URL. Knowing how to decode it on the spot turns a lot of opaque strings into readable data — which is exactly why a quick decoder is worth keeping a click away when you're inspecting a request or a config file.
Base64 is not compression — it costs size
A common misconception is that encoding "packs" data. It's the opposite: Base64 represents 3 bytes with 4 characters, so the output is about 33% larger than the input. That's a fine trade when you need binary to survive a text-only channel, but it's the reason you don't Base64 large files into a page — a big inlined image bloats the HTML and can't be cached separately. Use it for small assets and for moving bytes safely, not as a way to shrink anything.