URL Encoding Table — Percent Encoding Chart

Reserved, unreserved, and sub-delimiters

RFC 3986 splits URI characters into a few groups. Unreserved characters — letters, digits, and - _ . ~ — never need escaping anywhere in a URI, so every encoding mode leaves them alone. Reserved characters like : / ? # [ ] @ are structural: a / in a path is a segment separator, so a literal / that is meant as data, not structure, has to be escaped to %2F or it will be read as a boundary. Sub-delimiters like ! $ & ' ( ) * + , ; = carry meaning inside a specific component, most often a query string, where & separates parameters and = separates a key from its value.

Why the four columns differ

encodeURIComponent is built for encoding a single value that will sit inside a larger URI — a query parameter, for example — so it escapes every reserved and sub-delimiter character except ! ' ( ) *. encodeURI is built for encoding a whole URI that may already contain structural characters, so it leaves the reserved set untouched and only escapes characters that are never valid anywhere in a URI. Form encoding follows application/x-www-form-urlencoded, the rules a browser uses to submit an HTML form, which is why it turns a space into + rather than %20. Strict RFC 3986 is the specification's own minimal safe set, with no exceptions for the five sub-delimiters encodeURIComponent leaves alone — which is exactly why signature schemes like OAuth 1.0 and AWS SigV4 reject a naive encodeURIComponent call. The URL Encoder / Decoder names all four modes for the situation they fit, not just the function that implements them.

How this table is generated

Every value here comes from calling the exact same encoding function the URL Encoder / Decoder uses, so this table cannot drift out of sync with the tool. It covers every printable ASCII character plus a small, curated set of common non-ASCII characters — not a full sweep of Unicode, since a page per character is exactly the doorway-page shape this site avoids. Pasting a whole string instead of a single character? Use the encoder directly, or the URL Parser to split a full URL into its parts.