Magic Decoder

What you can do

• Decode a string without knowing what encoding was used.
• Unwrap multiple layers in one go, for example URL encoding over Base64 over gzip.
• Decode every common BaseXX: Base16, Base32, Base32hex, Base36, Base45, Base58, Base62, Base64 and Base85 (Ascii85 and Z85).
• Fix Base64 with missing padding or the URL-safe alphabet automatically.
• Read hex and binary dumps with prefixes, separators and line breaks.
• Decompress gzip, deflate, Brotli and Zstandard payloads found inside the data.
• Spot ROT13 by checking whether the rotated text reads like English.
• Pretty-print results that turn out to be JSON.
• Preview binary results as an image, PDF, audio or video, or inspect them in a hex view.
• Download binary results with the right file extension.

How to use the Magic Decoder

1. 1Paste the encoded text, or press Try an example to see a three-layer chain unwind.
2. 2Read the decoding chain to see which encodings were detected and in what order.
3. 3Copy or download the decoded result, or an intermediate value from any step.

How the detection works

Some encodings announce themselves clearly: percent escapes, HTML entities, Base64 padding, the <~ ~> wrapper of Ascii85, or the magic bytes of gzip and Zstandard. Others are ambiguous, an eight-character word can be valid hex, Base58 and Base64 at the same time. The tool ranks all plausible readings by confidence and only accepts an ambiguous one when its output actually looks meaningful, readable text, a recognizable file format or another decodable layer. That is what keeps ordinary words from being decoded into garbage.

The alphabet narrows things down further: a digit 8 rules out standard Base32, a lowercase letter rules out Base45, a 0 or uppercase O rules out Base58. Brotli and raw deflate carry no signature at all, so they are tried speculatively on binary data and kept only when the result makes sense. ROT13 is detected by comparing how much the text reads like English before and after rotation.

When the result is binary

Not every chain ends in text. If the final bytes are an image, an archive or something else binary, the tool recognizes common file types from their magic bytes. Images, PDFs, audio and video render as a preview right on the page, other formats appear as a hex dump, and the download gets a fitting file name. For compressed files and stubborn cases the Data Decompressor gives you manual control, and the Base64 Encoder & Decoder tool covers plain single-layer jobs.

Honest limits

The tool decodes, it does not decrypt. Encrypted data has no structure to detect, so ciphertext stays ciphertext. Encodings with custom alphabets, Base91 or Morse are not covered yet, and for genuinely ambiguous short strings the most confident reading wins, which is occasionally not the one you meant. The number-style bases (Base36, Base58, Base62) are only applied when the decoded bytes make sense, so a Bitcoin address, which decodes to opaque binary, is deliberately left alone.

Credits

Open source does the heavy lifting in this tool. Thank you to:

• brotli-wasm

  Google's Brotli compression, compiled to WebAssembly.

• zstd-wasm

  A WebAssembly build of Zstandard, the fast modern compression format.

Everything this site builds on is listed on the credits page.