Not a developer? Go to



Applies a regular expression operation on the input. This filter requires two values: the first is the pattern and the second is the replacement. Learning regular expressions is outside the scope of this document, though it is likely helpful to know that Perl-style regular expressions are supported. RegExr is a tool for learning, testing, and sharing regular expressions; a desktop version also available.

Note that if you don’t need regular expression support, the replace modifier is simpler and faster.

Replace Modifier

As noted above, the replace modifier is a better choice when a simple text search and replace is needed. An example of an operation better handled by replace is stripping the “.html” extension from URLs. Simply, a regular expression is not needed to do this type of search:

<mt:EntryPermalink replace=".html","">


Strip a Bracketed Phrase

This would strip any bracketed phrase from the end of the entry title field:

<mt:EntryTitle regex_replace="/\s*\[.+?\]\s*$/","">

Use Back References

Back references can be used in the replacement, such as this example that prints the first three words (if available, but at least 1) in the Entry Title:

<mt:EntryTitle regex_replace="/^(\w{1,3}).*/","$1">


Regex modifiers are supported. The global modifier, for example, will find all occurrences of bracketed phrases and remove them:

<mt:EntryTitle regex_replace="/\s*\[.+?\]\s*$/g",""$>

Or, a case-insensitive search to strip the word “apple” from text (in other words, it would find “Apple” or “aPpLlE”):

<mt:EntryTitle regex_replace="/\s*apple\s*/i","fruit">

Capturing Text in Parentheses

If blog description is:

<strong>I am the walrus</strong> <em>(Koo Koo Ca Choo)</em>

Use this regex_replace to get the part in parenthesis:

<mt:BlogDescription regex_replace="/(.*)<em>\((.*)\)</em>(.*)/","<em>$2</em>">

Output will be:

<em>Koo Koo Ca Choo</em>

Handling Double Quotes

Movable Type can get confused if the pattern string contains double quotes. It is best to define the pattern and replacement strings as variables, and reference those variables as the regex_replace values.

For example, the old name of your company, ACME “Widgets” Inc., might be strewn across your entries. But now that your company is called Widget Masters Inc., you might want to automatically replace the name in the output till you can find the time to do a quick “Search & Replace” on the entries.

<mt:SetVarBlock name="pattern">/ACME "Widgets" Inc/g</mt:SetVarBlock>
<mt:SetVarBlock name="replacement">Widget Masters Inc</mt:SetVarBlock>
<mt:EntryBody regex_replace="$pattern","$replacement">

You can do multiple regex replacements using different variable names. Note the modifiers are processed in order from left to right.

<mt:SetVarBlock name="p1">/ACME "Widgets" Inc/g</mt:SetVarBlock>
<mt:SetVarBlock name="r1">Widget Masters Inc</mt:SetVarBlock>
<mt:SetVarBlock name="p1">/AWI/g</mt:SetVarBlock>
<mt:SetVarBlock name="r1">WMI</mt:SetVarBlock>
<mt:EntryBody regex_replace="$p1","$r1" regex_replace="$p2","$r2">

2 Comments on May 16, 2014, 2:57 p.m. 返信

How can I use a positive lookbehind in the pattern string? MT sees the < in the (?<= operator and apparently treats it like an opening tag, causing nearby tags to break and the template to report an error.

Charlie Gorichanaz

Charlie Gorichanaz on May 19, 2014, 3:42 a.m. 返信

Hi there, that is a great question!

Any time you have trouble parsing values of template tag modifiers, such as regex_replace, you can first place the entire value inside a variable using SetVarBlock to avoid the issue.

For example, let’s say you want to replace letter E with # unless the E comes right after a P. You can achieve the positive lookbehind like this:

<mt:SetVarBlock name="p">/(?<!p)e/</mt:SetVarBlock>
<mt:SetVarBlock name="r">#</mt:SetVarBlock>
<mt:SetVarBlock name="original">Movable Type</mt:SetVarBlock>

Original: <$mt:Var name="original"$>
Replaced: <$mt:Var name="original" regex_replace="$p","$r"$>

The output is:

Original: Movable Type
Replaced: Movabl# Type