Regex Tester

Common Regex Patterns

• \d+ — one or more digits
• \w+ — one or more word characters (letters, digits, underscore)
• [a-zA-Z]+ — letters only
• ^\s+|\s+$ — leading or trailing whitespace
• (\w+)@(\w+.\w+) — simple email capture
• https?://\S+ — URLs

Tested Patterns

Each row shows what the tester returns with the g flag on:

A regex checks the format only. ^\d{4}-(?:0[1-9]|1[0-2])-(?:0[1-9]|[12]\d|3[01])$ accepts 2026-02-31, so parse dates with a date library.

Flags in Practice

The same text gives different results as flags change. With this three-line log as the test string:

ERROR disk full
WARN retry 3
ERROR timeout

• ^ERROR .+ with only g finds 1 match (ERROR disk full at index 0): ^ means the start of the whole text.
• Add m and it finds 2 matches, the second at index 29, because ^ now matches after every line break. The index counts the two newline characters too.
• . never crosses a line break unless s is on: a.b finds nothing in “a”, newline, “b” until you tick s.
• cafe with i on Café CAFE cafe finds CAFE and cafe but not Café; case-insensitive matching does not ignore accents.

Each match line shows its index, and each capture group is listed under it in order, including groups that did not take part in the match (shown as undefined). There is no Run button: the result updates 0.3 seconds after you stop typing. To try a pattern against a JSON document instead, the JSONPath tester queries by path rather than by text.

Common Mistakes

• Unescaped metacharacters. . * + ? ( ) [ ] { } ^ $ | \ have special meanings. Write \. to match a dot.
• ^ and $ in multi-line text. Without the m flag they match only at the start and end of the whole text.
• \w is ASCII only. \w+ on café 東京 matches only caf. Use a character range such as [\u4e00-\u9fff]+ for Chinese characters.
• Nested quantifiers. (a+)+$ can take a long time on a near-match. Matching runs in a native JavaScript RegExp Worker with a two-second total limit. Cancel stops it immediately. A timeout or cancellation gives no complete result. Changing the input stops the previous run.

Limits

• JavaScript syntax only. Python’s (?P<name>...) gives “Invalid group”; use (?<name>...). Patterns written for Go (RE2), PCRE, or .NET can behave differently.
• Four flags. Only g, i, m, and s are available. There is no u flag, so \p{L} is read as the literal text p{L}, and y, d, and v are not available.
• Groups are listed by number. Named groups appear as Group 1, Group 2, and so on.
• Full counts, bounded preview. Completed runs count every match. Highlights cover the first 100,000 UTF-16 units and at most 1,000 matches. The list shows the first 100 matches and 20 groups per match; each value is limited to 1,000 units, with a total value budget of 100,000 units. Shortened values end in …. Indices use UTF-16 units, so an emoji before a match adds 2. Worker support is required.
• No replace. The tester finds matches; it does not run replace().