Skip to content

HTML Entity Decoder

Decode HTML named and numeric character references with text, attribute-value or semicolon-required rules. Preview and copy one decoding layer.

Use this result well

Inputs that matter
Text, the HTML Entities Encode/Decode choice and, for decoding, the Text, Attribute value or Require semicolons context.
Output to expect
Escaped &, <, > and quote characters, or text decoded from HTML named and numeric character references.
How it works
Encoding replaces five characters once. Decoding follows the selected character-reference rules for one layer; a chain can choose a context for each HTML decoding step.
  • Text and attribute rules can differ when a semicolon is missing. Require semicolons keeps incomplete references unchanged; it does not validate a whole document.
  • Copy returns only the output text. Review it before inserting it into another context; converting references does not sanitize a page or execute its markup.
  • Text input is limited to 2,000,000 UTF-16 units and each result to 20,000,000. Large conversions can be cancelled or retried.
  • Chains allow up to 12 steps. Intermediate previews are shortened; Copy retains the complete final result.
Was this tool helpful?

Related Tools

Tools you might need next

HTML Entity Encoder

Encode ampersands, angle brackets and quotes as HTML entities. Preview the converted text, copy it, and compare examples before using it in markup.

Encoding
Base64 Decoder

Strictly decode standard or URL-safe Base64 into UTF-8 text, with explicit malformed-input errors and an untrusted-output security boundary.

Encoding
URL Decoder

Decode one percent-encoded URI or form field component with explicit plus handling, strict malformed-input errors, and inert output.

Encoding

Reference & details

How it works

Choose the character-reference context

Text uses HTML text decoding rules. Attribute value applies the additional boundary rule for incomplete named references. Require semicolons preserves references without a terminating semicolon.

Named and numeric references

Names are matched with their original case. Decimal and hexadecimal references use HTML numeric replacement rules, including replacement characters and legacy character mappings.

One layer of text conversion

References in the original input are decoded once. Markup is kept as text and newly produced references are not decoded again during that operation. Each HTML decoding step in a chain can select its own context.

Updated: September 2026

Example Scenarios

Read common and mathematical HTML references in Text mode.

Input: &copy; &eacute; &NotEqualTilde;

→ © é ≂̸

Use Text mode to inspect how a numeric prefix is consumed without losing the following characters.

Input: A &#65oops; B

→ A Aoops; B

Decode one ampersand layer before deciding whether to decode again.

Input: &amp;lt;

→ &lt;

FAQ

The decoder supports the HTML named-reference set, including &copy;, &eacute; and &NotEqualTilde;. Names are case-sensitive. Some names also have a recognized form without the final semicolon.

Use Text for ordinary HTML text or Attribute value for a value extracted from an attribute. Attribute rules keep an incomplete named reference when a letter, digit or equals sign follows it. Require semicolons decodes only complete semicolon-terminated references.

HTML text rules read the valid numeric prefix and keep the following characters. Require semicolons leaves that incomplete reference unchanged. The converter does not silently discard the suffix.

HTML decoding replaces zero, surrogate code points and numbers beyond the Unicode range with the replacement character. Some legacy numeric values have HTML-specific replacements: &#128; becomes the euro sign.

No. Output remains text, and a single operation decodes one layer. For example, &amp;lt; becomes &lt;. Copy returns only the output value, including the final output of a chain. Document validation and sanitizing require separate tools.

About HTML Entity Decoder

Decode HTML character references while keeping the surrounding markup as text. Choose Text, Attribute value or Require semicolons to match your input. The page opens in Text mode and decodes one layer at a time.