Free tool No sign-up

Base64 Encoder / Decoder

Encode or decode Base64 — text, JSON, JWTs, and files up to 5 MB. Standard and URL-safe variants with full UTF-8 support and live preview.

base64encoderdecoderjwtdata-uri

Plain Text Input

Examples:

Base64 Output

⟶
Type or paste on the left and hit Encode
or use Upload to Base64-encode a file

DirectionVariant

Why use Base64?

🔐

Inspect tokens & headers

JWTs, OAuth state, HTTP Basic auth, and signed cookies all use Base64. Decoding them on the spot beats writing a one-off script every time you debug an auth flow.

🖼️

Embed binary in text

Base64 is the universal way to ship images, fonts, or PDFs through APIs and inline data: URIs. The 33% size overhead is a small price for "no extra HTTP request" rendering.

🌐

URL-safe transport

Standard Base64 contains + and / which break inside URLs and filenames. Base64URL swaps them for - and _ — same encoding, safe for query strings, JWT segments, and S3 keys.

🧰

No round-tripping data loss

Unlike URL encoding or HTML entities, Base64 is a true 1:1 transform on bytes — encode, transmit, decode, and you get back exactly what you started with, every time.

▶

How to use this tool

Pick Encode or Decode

Encode turns plain text or files into Base64. Decode reads any Base64 string (with or without padding, URL-safe or standard) back to text or bytes.

Type, paste, or upload

Drop in any UTF-8 text — emoji, accented characters, and CJK all encode safely. For binary, click Upload to Base64-encode a file up to 5 MB right in the browser.

Toggle URL-safe if needed

Standard Base64 uses +, /, =. URL-safe (Base64URL, RFC 4648 §5) swaps + → -, / → _, and drops trailing =. Use it for JWT, query strings, and filenames.

Copy the result

For text, click Copy. For uploaded files, Copy data: URI gives you a ready-to-paste data:image/png;base64,... string for inlining in CSS, HTML, or Markdown.

⇄

Standard vs URL-safe Base64

Input	Standard	URL-safe	Note
`A`	`QQ==`	`QQ`	`1 byte → 4 chars + padding`
`AB`	`QUI=`	`QUI`	`2 bytes → 4 chars + padding`
`ABC`	`QUJD`	`QUJD`	`3 bytes → 4 chars exactly`
`Hello!`	`SGVsbG8h`	`SGVsbG8h`	`No special chars — same`
`sub?>tile`	`c3ViPz50aWxl`	`c3ViPz50aWxl`	`No +/ chars — same`
`\xfb\xff`	`+/8=`	`-_8`	`+, /, = → -, _, stripped`
`JWT header`	`eyJhbGc...InR5cCI6Ikp`	`eyJhbGc...InR5cCI6Ikp`	`JWTs always Base64URL`

Base64 in practice

UTF-8 safe

// Browser — UTF-8 safe encode
const text = "Café · 🚀";
const b64 = btoa(
  String.fromCharCode(...new TextEncoder().encode(text))
);
// → "Q2Fmw6kgwrcg8J+agA=="

Decode JWT

// Decode a JWT payload (always Base64URL)
const [, payload] = jwt.split(".");
const json = atob(
  payload.replace(/-/g, "+").replace(/_/g, "/")
         .padEnd(payload.length + (4 - payload.length % 4) % 4, "=")
);
// → {"sub":"1234","name":"Jane","exp":...}

data: URI

/* Inline a tiny image with a data: URI */
.icon-check {
  background-image: url("data:image/svg+xml;base64,PHN2Zy4uLjwvc3ZnPg==");
  background-size: 16px;
}

/* Or in HTML */
<img src="data:image/png;base64,iVBORw0KGgo...AAAAElFTkSuQmCC" alt="" />

✦

Pro tips

Base64 is encoding, not encryption

It looks scrambled but anyone can decode it. Never put secrets, API keys, or PII in Base64 thinking it's "hidden" — use real encryption (AES-GCM, age) for that.

Mind the +33% size penalty

3 bytes → 4 chars means Base64 grows files by ~33% (more after gzip can recover). Inline data URIs only beat HTTP for tiny assets — typically under 4 kB.

Always specify charset for text

Base64 encodes bytes, not text. If a system sends you Base64 of UTF-16 or Latin-1, decoding as UTF-8 produces mojibake. Confirm the original charset before decoding.

Watch for missing padding

Strict decoders reject Base64 without trailing =. JWTs and Base64URL strip them — most tools (including this one) re-pad automatically, but raw atob() in browsers does not.