Regex to Match Data URIs

This regex matches data URIs as defined in RFC 2397. Here is the token-by-token breakdown:

data: — Matches the literal data: scheme prefix that identifies a data URI.

([a-zA-Z]+/[a-zA-Z0-9.+-]+)? — Optional capturing group 1 matches the MIME type. It consists of a type (letters only) followed by a forward slash and a subtype (letters, digits, dots, plus signs, hyphens). Examples include text/plain, image/png, application/json, and application/vnd.ms-excel. The entire MIME type is optional; if omitted, text/plain is assumed.

(?:;([a-zA-Z0-9=]+))* — Zero or more non-capturing groups matching MIME parameters preceded by semicolons. Capturing group 2 captures the parameter value. Parameters include charset specifications like charset=UTF-8.

(?:;base64)? — An optional non-capturing group matching the literal ;base64 flag. When present, it indicates the data portion is Base64-encoded rather than percent-encoded.

, — Matches the literal comma that separates the metadata from the actual data content.

([^\s]*) — Capturing group 3 matches the data content: zero or more non-whitespace characters. For base64 data URIs, this contains the Base64-encoded content. For plain data URIs, this contains percent-encoded text.

The g flag enables global matching. Data URIs are used to embed small files directly in HTML, CSS, or JavaScript without requiring separate HTTP requests. This pattern is useful for detecting embedded resources, content security policy analysis, and resource extraction.