Base64 Encoder & Decoder

How to encode and decode Base64

1. 1 At the top, choose whether you are working with text or a file. Text can be encoded or decoded. A file, such as an image or a PDF, can be encoded to Base64 or straight to a data URI.
2. 2 For text, choose Encode or Decode and enter your text. The result appears instantly as you type.
3. 3 Copy the result with the Copy button. The Swap button moves it into the input and reverses the direction, for example to double-check it.

What the tool handles on its own

• when decoding, it recognizes both standard Base64 and base64url and adds any missing = signs at the end
• it reads a whole data URI, including the file type, and ignores spaces and line breaks
• it recognizes images, PDFs and archives, shows a preview for images and offers a download
• if the text is not UTF-8, it tries Windows-1250, the older Central European Windows code page, and tells you
• when encoding text, it also shows the base64url version below the result for URLs and JWT
• everything is processed right in your browser, nothing is sent to a server

How to convert an image to Base64 and back

Images are the most common use for Base64. You can encode a small image and embed it right in HTML or CSS, with no separate file. And when you receive an image as a Base64 string, for example from an API, you can turn it back into a file here and save it.

How to read a JWT token

A JWT token looks like one long string, but it consists of three parts separated by dots. The first two are plain JSON encoded in base64url, so you can read them without any key. Paste each part into the tool separately:

eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjMiLCJuYW1lIjoiSmFuYSIsImV4cCI6MTc2NzIyNTYwMH0.signature

Header

{"alg":"HS256","typ":"JWT"}

the signing algorithm and the token type

Payload

{"sub":"123","name":"Jana","exp":1767225600}

the user ID, name and expiry time

Signature

signature

a binary signature that the server verifies, not text

The exp value is the expiry time in seconds since January 1, 1970, here exactly midnight UTC on January 1, 2026. Anyone who has the token can read its payload, so never put passwords or other secrets in it.

How to extract an attachment from an email source

In the raw source of a message, every attachment is stored as a long block of Base64 after the Content-Transfer-Encoding: base64 header. In Gmail, open the source with Show original, in Outlook with View message source. Copy the block without the headers, paste it here, choose Decode and download the file. This helps when an attachment will not open in your email client or the client has blocked it.

Random key in Base64

Secret keys for APIs, encryption and signatures are often written as random bytes in Base64. Laravel, for example, stores its application key as APP_KEY=base64: followed by 32 random bytes. You can generate a key here too, using a cryptographically secure generator right in your browser.

What Base64 is and what it is for

Base64 is a way to write any data, even an image or a PDF, using just 64 common characters: uppercase and lowercase letters, digits and the + and / signs. Data written this way can travel anywhere only text is allowed, such as email attachments, JSON API responses, HTML or CSS. Every three bytes become four characters, so the result is about a third longer than the original data.

characters in the Base64 alphabet

64

every three bytes become four characters

3 → 4

how much longer the output gets

+33%

the standard that defines Base64

RFC 4648

Examples

This is how common text looks before and after encoding:

How Base64 works, step by step

1. 1 First, the text is converted to bytes. In UTF-8, plain English letters take one byte, accented letters such as é or č take two, and emoji take four.
2. 2 Each byte has eight bits. The bits are lined up one after another and split into groups of six.
3. 3 Each group of six bits is a number from 0 to 63, which maps to one character of the Base64 alphabet: 0 is A, 26 is a, 52 is 0, 62 is + and 63 is /.
4. 4 If there are not enough bits left to fill the last group, it is padded with zeros. Then = characters are added to the end so the length is divisible by four.

The Base64 alphabet (character table)

Base64 uses exactly 64 characters. Each one stands for a number from 0 to 63, which is one group of six bits. The = sign is not part of the alphabet. It only pads the length at the end.

1. 0 A
2. 1 B
3. 2 C
4. 3 D
5. 4 E
6. 5 F
7. 6 G
8. 7 H
9. 8 I
10. 9 J
11. 10 K
12. 11 L
13. 12 M
14. 13 N
15. 14 O
16. 15 P
17. 16 Q
18. 17 R
19. 18 S
20. 19 T
21. 20 U
22. 21 V
23. 22 W
24. 23 X
25. 24 Y
26. 25 Z
27. 26 a
28. 27 b
29. 28 c
30. 29 d
31. 30 e
32. 31 f
33. 32 g
34. 33 h
35. 34 i
36. 35 j
37. 36 k
38. 37 l
39. 38 m
40. 39 n
41. 40 o
42. 41 p
43. 42 q
44. 43 r
45. 44 s
46. 45 t
47. 46 u
48. 47 v
49. 48 w
50. 49 x
51. 50 y
52. 51 z
53. 52 0
54. 53 1
55. 54 2
56. 55 3
57. 56 4
58. 57 5
59. 58 6
60. 59 7
61. 60 8
62. 61 9
63. 62 +
64. 63 /

Why Base64 ends with = or ==

Base64 works on data three bytes at a time. When fewer bytes are left at the end, the = sign fills the missing space. One leftover byte adds two = signs, two bytes add one, and three bytes add none. Some systems, such as JWT, leave out the = signs because the decoder can work out the length on its own.

How big will the Base64 be

Every three bytes become four characters, so the Base64 output is 4 × ⌈n ÷ 3⌉ characters long, where n is the number of bytes and the brackets mean rounding up. That makes it roughly 33% longer. In emails, Base64 is also wrapped into lines of 76 characters, each ending with CR LF, so an attachment grows by about 37%. Going the other way, a Base64 string of length d holds 3 × d ÷ 4 bytes, minus the number of = signs at the end.

Base64 variants

All variants share the same first 62 characters. They differ in the last two characters, in padding and in whether the output is wrapped into lines.

Where Base64 is used in practice

<img src="data:image/png;base64,iVBORw0KGgo…" alt="Logo">

.icon {
  background: url("data:image/svg+xml;base64,PHN2Zy…");
}

Content-Type: application/pdf; name="invoice.pdf"
Content-Transfer-Encoding: base64

JVBERi0xLjcKJeLjz9MK…

Authorization: Basic dXNlcjpwYXNzd29yZA==

eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjMiLCJuYW1lIjoiSmFuYSJ9.signature

-----BEGIN CERTIFICATE-----
MIIDdzCCAl+gAwIBAgIE…
-----END CERTIFICATE-----

{
  "name": "contract.pdf",
  "content": "JVBERi0xLjcKJeLjz9MK…"
}

apiVersion: v1
kind: Secret
data:
  password: dGFqbmVoZXNsbw==

How to identify a file from the start of its Base64

Most formats start with typical bytes, so their Base64 always starts the same way too. The first few characters often tell you the file type before you even decode it. Rows marked with an asterisk only show the most common start, not a fixed rule.

Base64 compared with other encodings

Base64 is not the only way to write bytes as text. Other encodings are either longer but easier to read and type, or shorter but use less common characters.

A short history of Base64

1. 1987

   The encoding first appeared in the Privacy-Enhanced Mail (PEM) specification, RFC 989, under the name printable encoding.

2. 1992

   The MIME standard (RFC 1341) adopted it for email attachments, introduced 76-character lines and gave it the name Base64.

3. 2003

   RFC 3548 described Base64 on its own, outside of email, and added a variant for URLs and file names.

4. 2006

   RFC 4648 replaced the previous document and is still in force today.

Base64 in programming languages and on the command line

Here is how to encode and decode the text "Hello" in common languages. Always convert text with accented or other non-ASCII characters to UTF-8 bytes first, or you will get an error or a different result.

const base64 = btoa(String.fromCharCode(...new TextEncoder().encode('Hello')));
const text = new TextDecoder().decode(Uint8Array.from(atob(base64), (c) => c.charCodeAt(0)));

Buffer.from('Hello', 'utf8').toString('base64');
Buffer.from('SGVsbG8=', 'base64').toString('utf8');
Buffer.from('Hello').toString('base64url');

base64_encode('Hello');
base64_decode('SGVsbG8=', true);

import base64

base64.b64encode('Hello'.encode()).decode()
base64.b64decode('SGVsbG8=').decode()
base64.urlsafe_b64encode(b'Hello')

Base64.getEncoder().encodeToString("Hello".getBytes(StandardCharsets.UTF_8));
new String(Base64.getDecoder().decode("SGVsbG8="), StandardCharsets.UTF_8);

Convert.ToBase64String(Encoding.UTF8.GetBytes("Hello"));
Encoding.UTF8.GetString(Convert.FromBase64String("SGVsbG8="));

printf 'Hello' | base64
echo 'SGVsbG8=' | base64 -d
base64 < file.pdf > file.txt

[Convert]::ToBase64String([Text.Encoding]::UTF8.GetBytes('Hello'))
[Text.Encoding]::UTF8.GetString([Convert]::FromBase64String('SGVsbG8='))
[Convert]::ToBase64String([IO.File]::ReadAllBytes('file.pdf'))

Images and SVG in code

Image to Base64 in JavaScript

const file = document.querySelector('input[type=file]').files[0];
const reader = new FileReader();

reader.onload = () => console.log(reader.result);
reader.readAsDataURL(file);

The readAsDataURL method returns a complete data URI including the file type, for example data:image/png;base64,iVBORw0KGg….

Decoding a Base64 file on the command line

When you have Base64 saved in a file, for example from an API or a log, you can decode it straight back to the original file:

Linux

base64 -d image.txt > image.png

macOS

base64 -d -i image.txt -o image.png

Windows (Command Prompt)

certutil -decode image.txt image.png

Windows PowerShell

[IO.File]::WriteAllBytes('image.png', [Convert]::FromBase64String((Get-Content image.txt -Raw)))

OpenSSL

openssl base64 -d -A -in image.txt -out image.png

If the file starts with a data URI header such as data:image/png;base64,, delete it first. The certutil command also handles files with BEGIN and END lines.

Common errors and how to fix them

When to use Base64 and when not to