tiny-decode
Decoding HTML named character references traditionally requires a large table to store entity names and the code points to which they refer. Unless you take extreme care, this data can easily end up in your JavaScript bundle.
tiny-decode takes advantage of export conditions to include only as much JavaScript as necessary. For bundled browser code, the decode function is only a few bytes. For server-side code, decode includes the full entities module, which is much larger.
Example
Bundle `entities` 18.52 kB │ gzip: 19.64 kB
Bundle `tiny-decode` 0.97 kB │ gzip: 0.56 kB