Regex Replace

1. Paste source text. Include values that must change and near-matches that must remain untouched.
2. Define the find pattern. Choose a preset or enter a focused PCRE expression.
3. Write the replacement. Enter literal text or combine it with numbered captures.
4. Control scope. Select flags and use the limit when only the first few matches should change.
5. Audit the result. Compare highlighted source matches, positions, replacement count, warnings, and final text.

A replacement workflow, not a match debugger

Regex Replace answers a specific question: what will the complete text become after the selected matches are rewritten? The workspace first detects source matches, records their positions and captures, and then applies the replacement. A pattern can match correctly while the replacement is still wrong, so both stages are shown.

The summary distinguishes matches found from replacements made. These numbers differ when a replacement limit stops early. Character delta shows whether the result became shorter or longer, while the changed state identifies a replacement that produced text identical to the source.

Regex find-and-replace examples

Convert an ISO date to day/month/year

Pattern: (\d{4})-(\d{2})-(\d{2})
Replacement: $3/$2/$1
Input: Released 2026-09-04
Output: Released 04/09/2026

Collapse repeated horizontal spaces

Pattern: [ \t]{2,}
Replacement: single space
Input: alpha    beta
Output: alpha beta

Remove log-level prefixes per line

Pattern: ^(?:DEBUG|INFO|WARN|ERROR)\s*[:|-]\s*
Flags: mu
Replacement: empty

Swap comma-separated names

Pattern: ^([^,]+),\s*(.+)$
Flags: mu
Replacement: $2 $1
Input: Lovelace, Ada
Output: Ada Lovelace

Wrap order identifiers

Pattern: \b([A-Z]{2,5}-\d{3,8})\b
Replacement: [$1]
Input: Order API-2048 accepted
Output: Order [API-2048] accepted

Numbered capture references

A replacement reference uses the value captured by a pair of parentheses. If a date pattern contains year, month, and day groups, $3/$2/$1 reverses their order without retyping the values. Adding a new capturing group changes later numbers, so verify every reference after editing the pattern.

Use a non-capturing group such as (?:DEBUG|INFO) when parentheses are needed only to organize alternatives. This preserves the numbering of groups that the replacement actually uses.

Control the scope of every change

Anchors and boundaries prevent a transformation from spilling into nearby text. Multiline mode changes ^ and $ so they operate per line. Dot-all mode lets a dot cross line breaks. Case-insensitive mode can increase the set of affected values. Enable flags deliberately rather than copying them from an unrelated pattern.

Start with a short representative sample, use a limit of one when exploring an unfamiliar expression, and inspect every highlighted source match. Then test empty fields, unexpected punctuation, Unicode characters, and long nonmatching text.

Replacement safety and engine differences

Keep an untouched copy of important data. Regular expressions are not full parsers for HTML, JSON, XML, SQL, or programming languages. Use format-aware tools for structural edits and reserve regex for bounded text rules.

This workspace uses PCRE replacement syntax. Other languages can use different escaping or named-reference formats. Verify the final expression and replacement string in the destination runtime. See the PHP preg_replace reference and PCRE2 documentation.

Related Trexmi tools

Diagnose a pattern with Regex Tester PRO, generate a starting expression with Regex Builder, or collect matching values without modifying text with Regex Extractor.