HTML entity encoder and decoder

How to use

1. Choose "Encode" or "Decode".
2. For encoding, pick the format (named, decimal or hex) and which characters to encode: only the HTML special characters, or those plus every non-ASCII character.
3. Enter your text and press "Encode" or "Decode". Decoding reads named, decimal and hex references even when they are mixed.
4. Copy the result. "Use result as input" lets you convert the other way. The table at the bottom lists common references.

Notes and limits

• Input up to 2 MB.
• In named format, characters that have no name (emoji and others) are written as decimal references.
• Unknown names (such as &foo;) and numeric references to 0, surrogates or values above U+10FFFF are left unchanged, and the tool tells you how many there were.
• Decoding follows how browsers read references in page text. Inside attribute values, browsers treat some references without a semicolon differently.
• The result is always shown as plain text, never rendered as HTML. This tool does not strip tags or sanitize HTML.

How it works

The list of named references is generated from entities.json published with the WHATWG HTML Living Standard (2,231 entries).

Decoding never hands your input to the browser's HTML parser; a small parser reads each reference. A few legacy names such as &amp and &copy are recognized without the semicolon, and numeric references like &#65 may omit it too.

Numeric references 128 to 159 are read as Windows-1252 characters, as browsers do (for example – becomes an en dash).

Characters above U+FFFF, such as emoji, are encoded as one code point (😀), never as two surrogate halves.