Base64 Decode

How to decode Base64

1. Paste the Base64 text into the left box. A full data:image/png;base64,… URI works too, and so does Base64 that is split over many lines.
2. The decoded result appears on the right as you type. Text is read as UTF-8, so accents and emoji decode correctly.
3. If the data is a file, you get its type, a preview for images and a Download button with the right extension (.png, .pdf…).
4. If something is wrong, the message names the invalid character and its position. Press Select it to jump to it.

Standard Base64 (+ and /) and URL-safe Base64 (- and _) are both detected automatically, and missing = padding is added for you. To go the other way, use the Base64 encoder.

Base64 decode examples

These results come from the same code as the decoder above:

And two inputs that can't be decoded:

• SGVsbG8!IFdvcmxk: “Invalid character "!" at position 8.” The exclamation mark is not part of the Base64 alphabet.
• Zm9vY: “Base64 length is invalid: 5 characters leaves 1 extra character.” Five characters can't be a whole number of bytes, so part of the text is usually missing.

How Base64 decoding works

Base64 writes binary data with 64 safe characters. Each character stands for 6 bits, so 4 characters carry 24 bits, which is exactly 3 bytes. To decode Base64 manually:

1. Look up each character's value: A–Z = 0–25, a–z = 26–51, 0–9 = 52–61, + = 62, / = 63.
2. Write each value as 6 bits and join them.
3. Cut the bits into groups of 8. Each group is one byte. Leftover bits at the end (marked by = padding) are dropped.

Worked example, SGk=:

Joined: 01001000 01101001 00. The two full bytes are 72 and 105, which are “Hi” in ASCII; the last two zero bits are padding. Because 3 bytes become 4 characters, Base64 is about 33% larger than the data it carries.

Base64 to image, PDF or file

When the decoded bytes aren't text, the decoder identifies the file from its first bytes (its “magic number”) rather than trusting a label. You can often recognise a file from the start of its Base64:

For example, this 92-character string decodes to a 68 bytes PNG image (a single transparent pixel): iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAQAAAC1HAwCAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=. And JVBERi0xLjcK is the start of a PDF document. Images, including SVG, are previewed with an ordinary image tag, so scripts inside an SVG never run. PDFs, ZIP files and other types are offered as a download: check files from unknown sources before you open them.

How to decode a Base64 link or URL parameter

Base64 in links is usually URL-safe Base64 (also called base64url): + becomes -, / becomes _, and the = padding is often dropped, because +, / and = have special meanings in URLs. Paste just the encoded part, for example the value after ?data=. If you see %2B, %2F or %3D, the Base64 was also percent-encoded: decode it with the URL decoder first. A token made of three parts separated by dots is a JSON Web Token; the JWT decoder splits it and explains the claims.

Is Base64 encryption?

No. Base64 is an encoding: a reversible way to write bytes as text, with no key and no secret. Anyone who has the Base64 can decode it, exactly as this page does. That makes it safe for moving data around (email attachments, images in HTML or CSS, binary fields in JSON) but useless for hiding it. If you find passwords or API keys “protected” with Base64, treat them as exposed. Use real encryption (AES, or TLS in transit) for secrets, and a password hash such as bcrypt or Argon2 for passwords.

Decode Base64 in Python, JavaScript and the command line

Python (the base64 module):

import base64
base64.b64decode("SGVsbG8sIFdvcmxkIQ==").decode("utf-8")  # 'Hello, World!'
s = "eyJ1c2VyIjoiYWRhIiwicm9sZSI6ImFkbWluIn0"
base64.urlsafe_b64decode(s + "=" * (-len(s) % 4))          # URL-safe, padding restored

JavaScript in the browser. atob() alone returns one character per byte, which garbles UTF-8 (“café” comes out as “café”), so decode the bytes with TextDecoder:

const bytes = Uint8Array.from(atob(b64), (c) => c.charCodeAt(0));
const text = new TextDecoder().decode(bytes);

Node.js: Buffer.from(b64, "base64").toString("utf8") (it also accepts URL-safe input).

Linux and macOS: echo 'SGVsbG8sIFdvcmxkIQ==' | base64 -d, or base64 -d image.txt > image.png for a file (older macOS versions use -D). Windows PowerShell: [Text.Encoding]::UTF8.GetString([Convert]::FromBase64String("SGVsbG8sIFdvcmxkIQ==")).

Related developer tools

• Base64 encoder: text, images and files to Base64 or a data URI.
• URL decoder: turn %20, %3D and other percent-escapes back into text.
• JWT decoder: read a token's header, payload and expiry, and verify its signature.
• JSON formatter: pretty-print JSON you decoded from Base64.

Frequently asked questions