Quick answer
Base64 represents any data with 64 safe characters (A–Z, a–z, 0–9, + and /), turning every 3 bytes into 4 characters, so output is about 33% larger. “Hello, World!” (13 bytes) encodes to SGVsbG8sIFdvcmxkIQ==.
What is Base64?
This tool encodes and decodes Base64 for UTF-8 text and for any file up to 10 MB. It supports standard Base64, Base64URL as used in JWTs, 76-character MIME line wrapping and data URIs; when decoding back to a file, the file type is detected from its content. Everything runs in your browser.
How to encode and decode Base64
- 1
“Text” tab: choose “Encode → Base64” or “Decode Base64 →”, then type or paste; the result appears alongside.
- 2
Turn on “Base64URL” for URLs, file names or JWTs, and line wrapping for email. Click “Swap” to decode the result back.
- 3
“File” tab → “File → Base64”: click “Choose a file to encode”, optionally output a data URI, then copy or “Download .txt”.
- 4
“File” tab → “Base64 → File”: paste Base64 or a data URI, check the detected type, preview images and download the file.
Base64 variants
| Variant | Characters | Trailing = | Used for |
|---|---|---|---|
| Standard Base64 (RFC 4648) | A–Z a–z 0–9 + / | Yes | Data in JSON, XML, configs |
| Base64URL (RFC 4648 §5) | A–Z a–z 0–9 - _ | Usually dropped (this tool drops it) | URLs, file names, JWTs |
| MIME (email) | Same as standard | Yes | Email attachments, lines wrapped every 76 characters |
| Data URI | Standard, with a data:…;base64, prefix | Yes | Embedding small images or fonts in HTML / CSS |
Swipe sideways to see the whole table →
Size after encoding
Every 3 bytes become 4 Base64 characters (rounded up), so the output is about 4/3 of the original.
| Original | Base64 | Notes |
|---|---|---|
| 1 byte | 4 characters (two =) | Padding fills the group of four |
| 13 bytes (“Hello, World!”) | 20 characters | Accented letters take 2–3 bytes in UTF-8 |
| 300 KB | About 400 KB | Images embedded as data URIs make HTML heavier |
| 3 MB | 4 MB | Wrapping every 76 characters adds about 1.3% in line breaks |
Real-world examples
Reading a JWT payload
A JWT has three dot-separated parts. Paste the middle one in decode mode: Base64URL is detected and the JSON payload (user, expiry…) is shown. The signature is not verified.
Inlining a small icon in CSS
Pick a small SVG or PNG on the File tab, enable data URI and paste it into background-image: url(…). Only worth it for tiny files, since size grows by about a third.
Recovering an image from an API
An API returns an image as a Base64 string? Paste it into “Base64 → File”: PNG or JPEG is detected from the content, previewed and downloaded with the right extension.
Encoding accented text
“Xin chào Việt Nam” becomes WGluIGNow6BvIFZp4buHdCBOYW0= and decodes back exactly, because UTF-8 is used in both directions.
Common mistakes
- Decoding gives garbage or “not UTF-8 text”.
- The original data is a binary file (image, PDF, zip…). Use the “File” tab → “Base64 → File” to restore the file instead of reading it as text.
- Invalid length error.
- Characters are missing, usually from an incomplete copy. Spaces and line breaks are fine; they are ignored.
- Character not in the Base64 alphabet.
- Stray spaces, brackets or surrounding text crept in. The message gives the exact position to delete.
- “=” in the wrong place.
- At most two = are allowed, at the very end. You may have joined two Base64 strings; decode each separately.
- Files over 10 MB cannot be chosen.
- The 10 MB cap keeps the browser responsive. For bigger files, send the original file rather than Base64.
- Treating Base64 as security.
- Base64 just rewrites data and anyone can decode it. Never use it to hide passwords or sensitive data.
Glossary
- Base64 encoding
- Turning bytes into a string of 64 safe characters so data can travel through text-only systems.
- Padding (=)
- One or two = at the end that complete the last group of four when data is not a multiple of 3 bytes.
- Base64URL
- A variant replacing + / with - _ so strings are safe in URLs and file names.
- Data URI
- A string like data:<type>;base64,<data> that embeds a file directly in HTML or CSS.
- UTF-8
- How text becomes bytes; accented letters take 2–3 bytes, which is why UTF-8 is needed to decode them correctly.
- Magic bytes
- The first few bytes that identify a file type (PNG, JPEG, PDF…), used to detect a decoded file.
Frequently asked questions
Is Base64 encryption?
No. Base64 only changes how data is written; there is no secret key and anyone can decode it. Use real encryption such as AES for secrecy.
Is Base64URL detected automatically when decoding?
Yes. Strings with - or _ or without trailing = decode fine, and the tool reports that Base64URL was detected.
Can I paste a data URI?
Yes. The “data:…;base64,” prefix is removed automatically; a data URI without “;base64” is reported clearly.
Which file types are detected?
More than 25 common ones such as PNG, JPEG, GIF, WebP, PDF, ZIP, SVG, MP3 and MP4, based on the first bytes. Unknown data still downloads with a .bin extension.
Will very long Base64 strings slow the page down?
Long results only show their beginning and large pasted input is collapsed; copying and downloading always use the full data.
Is my data uploaded?
No. Text and files are processed in your browser, including when you download the result.
Ratings and comments
Loading…