Why does my pattern work here but not in my code?

Regex syntax differs slightly between engines — JavaScript, Python, and PCRE (used by many other languages) handle things like named groups, lookbehind, and Unicode property escapes differently, so always confirm which flavor your target language uses.

What's the difference between a match and a capture group?

The match is the entire substring the pattern found; capture groups, created with parentheses, let you pull out specific sub-parts of that match, like extracting just the domain from a matched email address.

Why does my pattern match too much text?

Greedy quantifiers (*, +, {n,}) consume as many characters as possible by default; add a ? after the quantifier (like *?) to make it lazy and match the shortest possible string instead.

Can regex validate something complex like a fully RFC-compliant email address?

Technically only with an extremely long, hard-to-read pattern — most real-world validators use a simplified regex for basic format checks and rely on actually sending a confirmation email to verify it's real and deliverable.

tools / developer

Regex Tester

Test regular expressions with live highlighting.

TEST STRING

QUICK REFERENCE

.any character except newline

\ddigit [0-9]

\wword char [a-zA-Z0-9_]

\swhitespace

^start of string/line

$end of string/line

*0 or more

+1 or more

?0 or 1 (optional)

{n,m}between n and m times

(abc)capture group

(?:abc)non-capture group

a|ba or b

[abc]character class

[^abc]negated class

ABOUT THIS TOOL

Write a regular expression and paste sample text to see matches highlighted in real time as you type. Supports standard flags like global (g), case-insensitive (i), and multiline (m), and shows capture groups separately from the full match so you can see exactly what a pattern extracts versus what it simply matches. A practical way to build, test, and refine patterns before dropping them into code, since regex behavior can vary subtly between languages — JavaScript, Python, and PCRE handle lookbehind, Unicode property escapes, and named groups differently — and testing against real sample text catches edge cases no amount of trial and error in production would.

HOW TO USE

Type your regular expression pattern into the pattern field.
Set any flags you need — global to find all matches, case-insensitive to ignore letter case, multiline for ^ and $ to match per line.
Paste representative sample text into the test area, including edge cases you're worried about.
Watch matches highlight live as you edit the pattern.
Inspect capture groups separately from the overall match to confirm what your pattern would extract in code.
Refine the pattern until it matches everything it should and nothing it shouldn't, then copy it into your codebase.

COMMON USE CASES

A developer writing a form validator builds a pattern to check whether an email or phone number format is valid before submitting.
Someone parsing log files crafts a regex to extract IP addresses or timestamps from unstructured text.
A developer doing a find-and-replace across a codebase tests a pattern against sample lines before running it in their editor's regex search.
A data engineer cleaning a messy CSV writes a pattern to strip unwanted characters or split combined fields.
Someone debugging why a regex isn't matching in production pastes the exact failing input to see which part of the pattern is the problem.

TIPS & COMMON MISTAKES

Greedy quantifiers like .* match as much as possible and can overshoot across unexpected text — try the lazy version .*? when you want the shortest possible match instead.
Regex flavors differ: JavaScript only added lookbehind assertions in newer engines and lacks some PCRE features, so a pattern written for Python's re module may need tweaking to run in JavaScript.
Special characters like ., *, +, ?, (, ), [, and \ have meaning in regex and must be escaped with a backslash if you want to match them literally.
Without the global flag, most regex methods stop after the first match — forgetting it is a common reason "it only replaced one instance" bugs happen.