Advanced Usage

1. Character Set Handling

A character set is basically a mapping between bytes and glyphs and implies a certain character encoding scheme. For example, for the ISO 8859 family of character sets, an encoding of 8bit per character is used. For the Unicode character set, different character encodings may be used, UTF-8 being the most popular. In UTF-8, a character is represented using a variable number of bytes ranging from 1 to 4.

Since Mutt is a command-line tool run from a shell, and delegates certain tasks to external tools (such as an editor for composing/editing messages), all of these tools need to agree on a character set and encoding. There exists no way to reliably deduce the character set a plain text file has. Interoperability is gained by the use of well-defined environment variables. The full set can be printed by issuing locale on the command line.

Upon startup, Mutt determines the character set on its own using routines that inspect locale-specific environment variables. Therefore, it is generally not necessary to set the $charset variable in Mutt. It may even be counter-productive as Mutt uses system and library functions that derive the character set themselves and on which Mutt has no influence. It's safest to let Mutt work out the locale setup itself.

If you happen to work with several character sets on a regular basis, it's highly advisable to use Unicode and an UTF-8 locale. Unicode can represent nearly all characters in a message at the same time. When not using a Unicode locale, it may happen that you receive messages with characters not representable in your locale. When displaying such a message, or replying to or forwarding it, information may get lost possibly rendering the message unusable (not only for you but also for the recipient, this breakage is not reversible as lost information cannot be guessed).

A Unicode locale makes all conversions superfluous which eliminates the risk of conversion errors. It also eliminates potentially wrong expectations about the character set between Mutt and external programs.

The terminal emulator used also must be properly configured for the current locale. Terminal emulators usually do not derive the locale from environment variables, they need to be configured separately. If the terminal is incorrectly configured, Mutt may display random and unexpected characters (question marks, octal codes, or just random glyphs), format strings may not work as expected, you may not be abled to enter non-ascii characters, and possible more. Data is always represented using bytes and so a correct setup is very important as to the machine, all character sets look the same.

Warning: A mismatch between what system and library functions think the locale is and what Mutt was told what the locale is may make it behave badly with non-ascii input: it will fail at seemingly random places. This warning is to be taken seriously since not only local mail handling may suffer: sent messages may carry wrong character set information the receiver has too deal with. The need to set $charset directly in most cases points at terminal and environment variable setup problems, not Mutt problems.

A list of officially assigned and known character sets can be found at IANA, a list of locally supported locales can be obtained by running locale -a.

2. Regular Expressions

All string patterns in Mutt including those in more complex patterns must be specified using regular expressions (regexp) in the POSIX extended syntax (which is more or less the syntax used by egrep and GNU awk). For your convenience, we have included below a brief description of this syntax.

The search is case sensitive if the pattern contains at least one upper case letter, and case insensitive otherwise.


\ must be quoted if used for a regular expression in an initialization command: \\.

A regular expression is a pattern that describes a set of strings. Regular expressions are constructed analogously to arithmetic expressions, by using various operators to combine smaller expressions.


The regular expression can be enclosed/delimited by either " or ' which is useful if the regular expression includes a white-space character. See Syntax of Initialization Files for more information on " and ' delimiter processing. To match a literal " or ' you must preface it with \ (backslash).

The fundamental building blocks are the regular expressions that match a single character. Most characters, including all letters and digits, are regular expressions that match themselves. Any metacharacter with special meaning may be quoted by preceding it with a backslash.

The period . matches any single character. The caret ^ and the dollar sign $ are metacharacters that respectively match the empty string at the beginning and end of a line.

A list of characters enclosed by [ and ] matches any single character in that list; if the first character of the list is a caret ^ then it matches any character not in the list. For example, the regular expression [0123456789] matches any single digit. A range of ASCII characters may be specified by giving the first and last characters, separated by a hyphen -. Most metacharacters lose their special meaning inside lists. To include a literal ] place it first in the list. Similarly, to include a literal ^ place it anywhere but first. Finally, to include a literal hyphen - place it last.

Certain named classes of characters are predefined. Character classes consist of [:, a keyword denoting the class, and :]. The following classes are defined by the POSIX standard in Table 4.1, “POSIX regular expression character classes”

Table 4.1. POSIX regular expression character classes

Character class Description
[:alnum:] Alphanumeric characters
[:alpha:] Alphabetic characters
[:blank:] Space or tab characters
[:cntrl:] Control characters
[:digit:] Numeric characters
[:graph:] Characters that are both printable and visible. (A space is printable, but not visible, while an a is both)
[:lower:] Lower-case alphabetic characters
[:print:] Printable characters (characters that are not control characters)
[:punct:] Punctuation characters (characters that are not letter, digits, control characters, or space characters)
[:space:] Space characters (such as space, tab and formfeed, to name a few)
[:upper:] Upper-case alphabetic characters
[:xdigit:] Characters that are hexadecimal digits

A character class is only valid in a regular expression inside the brackets of a character list.


Note that the brackets in these class names are part of the symbolic names, and must be included in addition to the brackets delimiting the bracket list. For example, [[:digit:]] is equivalent to [0-9] .

Two additional special sequences can appear in character lists. These apply to non-ASCII character sets, which can have single symbols (called collating elements) that are represented with more than one character, as well as several characters that are equivalent for collating or sorting purposes:

Collating Symbols

A collating symbol is a multi-character collating element enclosed in [. and .]. For example, if ch is a collating element, then [[.ch.]] is a regexp that matches this collating element, while [ch] is a regexp that matches either c or h.

Equivalence Classes

An equivalence class is a locale-specific name for a list of characters that are equivalent. The name is enclosed in [= and =]. For example, the name e might be used to represent all of e with grave ( è), e with acute ( é) and e. In this case, [[=e=]] is a regexp that matches any of: e with grave ( è), e with acute ( é) and e.

A regular expression matching a single character may be followed by one of several repetition operators described in Table 4.2, “Regular expression repetition operators”.

Table 4.2. Regular expression repetition operators

Operator Description
? The preceding item is optional and matched at most once
* The preceding item will be matched zero or more times
+ The preceding item will be matched one or more times
{n} The preceding item is matched exactly n times
{n,} The preceding item is matched n or more times
{,m} The preceding item is matched at most m times
{n,m} The preceding item is matched at least n times, but no more than m times

Two regular expressions may be concatenated; the resulting regular expression matches any string formed by concatenating two substrings that respectively match the concatenated subexpressions.

Two regular expressions may be joined by the infix operator |; the resulting regular expression matches any string matching either subexpression.

Repetition takes precedence over concatenation, which in turn takes precedence over alternation. A whole subexpression may be enclosed in parentheses to override these precedence rules.


If you compile Mutt with the included regular expression engine, the following operators may also be used in regular expressions as described in Table 4.3, “GNU regular expression extensions”.

Table 4.3. GNU regular expression extensions

Expression Description
\\y Matches the empty string at either the beginning or the end of a word
\\B Matches the empty string within a word
\\< Matches the empty string at the beginning of a word
\\> Matches the empty string at the end of a word
\\w Matches any word-constituent character (letter, digit, or underscore)
\\W Matches any character that is not word-constituent
\\` Matches the empty string at the beginning of a buffer (string)
\\' Matches the empty string at the end of a buffer

Please note however that these operators are not defined by POSIX, so they may or may not be available in stock libraries on various systems.

3. Patterns: Searching, Limiting and Tagging

3.1. Pattern Modifier

Many of Mutt's commands allow you to specify a pattern to match ( limit, tag-pattern, delete-pattern, etc.). Table 4.4, “Pattern modifiers” shows several ways to select messages.

Table 4.4. Pattern modifiers

Pattern modifier Description
~A all messages
~b EXPR messages which contain EXPR in the message body
=b STRING messages which contain STRING in the message body. If IMAP is enabled, searches for STRING on the server, rather than downloading each message and searching it locally.
~B EXPR messages which contain EXPR in the whole message
=B STRING messages which contain STRING in the whole message. If IMAP is enabled, searches for STRING on the server, rather than downloading each message and searching it locally.
~c EXPR messages carbon-copied to EXPR
%c GROUP messages carbon-copied to any member of GROUP
~C EXPR messages either to: or cc: EXPR
%C GROUP messages either to: or cc: to any member of GROUP
~d [ MIN ]-[ MAX ] messages with date-sent in a Date range
~D deleted messages
~e EXPR messages which contains EXPR in the Sender field
%e GROUP messages which contain a member of GROUP in the Sender field
~E expired messages
~F flagged messages
~f EXPR messages originating from EXPR
%f GROUP messages originating from any member of GROUP
~g cryptographically signed messages
~G cryptographically encrypted messages
~h EXPR messages which contain EXPR in the message header
=h STRING messages which contain STRING in the message header. If IMAP is enabled, searches for STRING on the server, rather than downloading each message and searching it locally; STRING must be of the form header: substring(see below).
~H EXPR messages with a spam attribute matching EXPR
~i EXPR messages which match EXPR in the Message-ID field
~k messages which contain PGP key material
~L EXPR messages either originated or received by EXPR
%L GROUP message either originated or received by any member of GROUP
~l messages addressed to a known mailing list
~m [ MIN ]-[ MAX ] messages with numbers in the range MIN to MAX ***)
~m <[ MAX ] messages with numbers less than MAX ***)
~m >[ MIN ] messages with numbers greater than MIN ***)
~m [ M ] just message number M ***)
~m [ MIN ],[ MAX ] messages with offsets (from selected message) in the range MIN to MAX ***)
~n [ MIN ]-[ MAX ] messages with a score in the range MIN to MAX *)
~N new messages
~O old messages
~p messages addressed to you (consults alternates )
~P messages from you (consults alternates )
~Q messages which have been replied to
~r [ MIN ]-[ MAX ] messages with date-received in a Date range
~R read messages
~s EXPR messages having EXPR in the Subject field.
~S superseded messages
~t EXPR messages addressed to EXPR
~T tagged messages
~u messages addressed to a subscribed mailing list
~U unread messages
~v messages part of a collapsed thread.
~V cryptographically verified messages
~x EXPR messages which contain EXPR in the References or In-Reply-To field
~X [ MIN ]-[ MAX ] messages with MIN to MAX attachments *)
~y EXPR messages which contain EXPR in their keywords
~z [ MIN ]-[ MAX ] messages with a size in the range MIN to MAX *) **)
=/ STRING IMAP custom server-side search for STRING . Currently only defined for Gmail. See: GMail Patterns
~= duplicated messages (see $duplicate_threads)
~# broken threads (see $strict_threads)
~$ unreferenced messages (requires threaded view)
~( PATTERN ) messages in threads containing messages matching PATTERN , e.g. all threads containing messages from you: ~(~P)
~<( PATTERN ) messages whose immediate parent matches PATTERN , e.g. replies to your messages: ~<(~P)
~>( PATTERN ) messages having an immediate child matching PATTERN , e.g. messages you replied to: ~>(~P)

Where EXPR is a regular expression, and GROUP is an address group.

*) The forms <[ MAX ], >[ MIN ], [ MIN ]- and -[ MAX ] are allowed, too.

**) The suffixes K and M are allowed to specify kilobyte and megabyte respectively.

***) The message number ranges (introduced by ~m) are even more general and powerful than the other types of ranges. Read on and see Section 3.1.1, “Message Ranges” below.

Special attention has to be paid when using regular expressions inside of patterns. Specifically, Mutt's parser for these patterns will strip one level of backslash ( \), which is normally used for quoting. If it is your intention to use a backslash in the regular expression, you will need to use two backslashes instead ( \\). You can force Mutt to treat EXPR as a simple substring instead of a regular expression by using = instead of ~ in the pattern name. For example, =b *.*will find all messages that contain the literal string *.*. Simple string matches are less powerful than regular expressions but can be considerably faster. This is especially true for IMAP folders, because string matches can be performed on the server instead of by fetching every message. IMAP treats =h specially: it must be of the form header: substring and will not partially match header names. The substring part may be omitted if you simply wish to find messages containing a particular header without regard to its value.

Patterns matching lists of addresses (notably c, C, p, P and t) match if there is at least one match in the whole list. If you want to make sure that all elements of that list match, you need to prefix your pattern with ^. This example matches all mails which only has recipients from Germany.

Example 4.1. Matching all addresses in address lists

^~C \.de$

You can restrict address pattern matching to aliases that you have defined with the "@" modifier. This example matches messages whose recipients are all from Germany, and who are known to your alias list.

Example 4.2. Matching restricted to aliases

^@~C \.de$

To match any defined alias, use a regular expression that matches any string. This example matches messages whose senders are known aliases.

Example 4.3. Matching any defined alias

@~f .

3.1.1. Message Ranges

If a message number range (from now on: MNR) contains a comma ( ,), it is a relative MNR. That means the numbers denote offsets from the highlighted message. For example:

Table 4.5. Relative Message Number Ranges

Pattern Explanation
~m -2,2 Previous 2, highlighted and next 2 emails
~m 0,1 Highlighted and next email

In addition to numbers, either side of the range can also contain one of the special characters (shortcuts) .^$. The meaning is:

Table 4.6. Message Number Shortcuts

Shortcut Explanation Example Meaning
. Current / Highlighted ~m -3,. Previous 3 emails plus the highlighted one
$ Last ~m .,$ Highlighted email and all the later ones
^ First ~m ^,1 Highlighted, next and all preceding ones

Lastly, you can also leave either side of the range blank, to make it extend as far as possible. For example, ~m ,1 has the same meaning as the last example in Table 4.6, “Message Number Shortcuts”.

Otherwise, if a MNR doesn't contain a comma, the meaning is similar to other ranges, except that the shortcuts are still available. Examples:

Table 4.7. Absolute Message Number Ranges

Pattern Explanation
~m 3-10 Emails 3 to 10
~m -10 Emails 1 to 10
~m 10- Emails 10 to last
~m <3 First and second email
~m ^-2 First and second email
~m >1 Everything but first email
~m 2-$ Everything but first email
~m 2 Just the second email

3.2. Simple Searches

Mutt supports two versions of so called simple searches. These are issued if the query entered for searching, limiting and similar operations does not seem to contain a valid pattern modifier (i.e. it does not contain one of these characters: ~, = or %). If the query is supposed to contain one of these special characters, they must be escaped by prepending a backslash ( \).

The first type is by checking whether the query string equals a keyword case-insensitively from Table 4.8, “Simple search keywords”: If that is the case, Mutt will use the shown pattern modifier instead. If a keyword would conflict with your search keyword, you need to turn it into a regular expression to avoid matching the keyword table. For example, if you want to find all messages matching flag(using $simple_search) but don't want to match flagged messages, simply search for [f]lag .

Table 4.8. Simple search keywords

Keyword Pattern modifier
all ~A
. ~A
^ ~A
del ~D
flag ~F
new ~N
old ~O
repl ~Q
read ~R
tag ~T
unread ~U

The second type of simple search is to build a complex search pattern using $simple_search as a template. Mutt will insert your query properly quoted and search for the composed complex query.

3.3. Nesting and Boolean Operators

Logical AND is performed by specifying more than one criterion. For example:

~t mutt ~f elkins

would select messages which contain the word mutt in the list of recipients and that have the word elkins in the From header field.

Mutt also recognizes the following operators to create more complex search patterns:

  • ! — logical NOT operator

  • | — logical OR operator

  • () — logical grouping operator

Here is an example illustrating a complex search pattern. This pattern will select all messages which do not contain mutt in the To or Cc field and which are from elkins.

Example 4.4. Using boolean operators in patterns

!(~t mutt|~c mutt) ~f elkins

Here is an example using white space in the regular expression (note the ' and " delimiters). For this to match, the mail's subject must match the ^Junk +From +Me$ and it must be from either Jim +Somebody or Ed +SomeoneElse:

'~s "^Junk +From +Me$" ~f ("Jim +Somebody"|"Ed +SomeoneElse")'


If a regular expression contains parenthesis, or a vertical bar ("|"), you must enclose the expression in double or single quotes since those characters are also used to separate different parts of Mutt's pattern language. For example: ~f "me@(mutt\.org|cs\.hmc\.edu)"Without the quotes, the parenthesis wouldn't end. This would be separated to two OR'd patterns: ~f me@(mutt\.org and cs\.hmc\.edu) . They are never what you want.

3.4. Searching by Date

Mutt supports two types of dates, absolute and relative .

3.4.1. Absolute Dates

Dates must be in DD/MM/YY format (month and year are optional, defaulting to the current month and year). An example of a valid range of dates is:

Limit to messages matching: ~d 20/1/95-31/10

If you omit the minimum (first) date, and just specify -DD/MM/YY, all messages before the given date will be selected. If you omit the maximum (second) date, and specify DD/MM/YY-, all messages after the given date will be selected. If you specify a single date with no dash ( -), only messages sent on the given date will be selected.

You can add error margins to absolute dates. An error margin is a sign (+ or -), followed by a digit, followed by one of the units in Table 4.9, “Date units”. As a special case, you can replace the sign by a * character, which is equivalent to giving identical plus and minus error margins.

Table 4.9. Date units

Unit Description
y Years
m Months
w Weeks
d Days

Example: To select any messages two weeks around January 15, 2001, you'd use the following pattern:

Limit to messages matching: ~d 15/1/2001*2w

3.4.2. Relative Dates

This type of date is relative to the current date, and may be specified as:

  • > offset for messages older than offset units

  • < offset for messages newer than offset units

  • = offset for messages exactly offset units old

offset is specified as a positive number with one of the units from Table 4.9, “Date units”.

Example: to select messages less than 1 month old, you would use

Limit to messages matching: ~d <1m


All dates used when searching are relative to the local time zone, so unless you change the setting of your $index_format to include a %[...]format, these are not the dates shown in the main index.

3.5. GMail Patterns

=/ "search terms"invokes server-side search, passing along the search terms provided. Search results are constrained by IMAP to be within the current folder. At present this only supports Gmail's search API IMAP extension. The search language is entirely up to the mail provider and changes at their discretion. Using ~/will silently fail.

For up-to-date information about searching, see: GMail's Support Page. You will need to (once) use a web-browser to visit Settings/Labels and enable "Show in IMAP" for "All Mail". When searching, visit that folder in mutt to most closely match Gmail search semantics.

Table 4.10. GMail Example Patterns

Pattern Matches
=/ " has:attachment is:important" the mailing-list per Gmail's definitions, and has an attachment, and has been marked as important
=/ "{has:purple-star has:yellow-star} older_than:2m" is older than two months and has either a purple-star or a yellow-star

4. Marking Messages

There are times that it's useful to ask Mutt to "remember" which message you're currently looking at, while you move elsewhere in your mailbox. You can do this with the mark-message operator, which is bound to the ~ key by default. Press this key to enter an identifier for the marked message. When you want to return to this message, press ' and the name that you previously entered.

(Message marking is really just a shortcut for defining a macro that returns you to the current message by searching for its Message-ID. You can choose a different prefix by setting the $mark_macro_prefix variable.)

5. Using Tags

Sometimes it is desirable to perform an operation on a group of messages all at once rather than one at a time. An example might be to save messages to a mailing list to a separate folder, or to delete all messages with a given subject. To tag all messages matching a pattern, use the <tag-pattern>function, which is bound to shift-T by default. Or you can select individual messages by hand using the <tag-message>function, which is bound to t by default. See patterns for Mutt's pattern matching syntax.

Once you have tagged the desired messages, you can use the tag-prefix operator, which is the ;(semicolon) key by default. When the tag-prefix operator is used, the next operation will be applied to all tagged messages if that operation can be used in that manner. If the $auto_tag variable is set, the next operation applies to the tagged messages automatically, without requiring the tag-prefix.

In macro s or push commands, you can use the <tag-prefix-cond>operator. If there are no tagged messages, Mutt will eat the rest of the macro to abort it's execution. Mutt will stop eating the macro when it encounters the <end-cond>operator; after this operator the rest of the macro will be executed as normal.

6. Using Hooks

A hook is a concept found in many other programs which allows you to execute arbitrary commands before performing some operation. For example, you may wish to tailor your configuration based upon which mailbox you are reading, or to whom you are sending mail. In the Mutt world, a hook consists of a regular expression or pattern along with a configuration option/command. See:

for specific details on each type of hook available.


If a hook changes configuration settings, these changes remain effective until the end of the current Mutt session. As this is generally not desired, a default hook needs to be added before all other hooks of that type to restore configuration defaults.

Example 4.5. Specifying a default hook

send-hook . 'unmy_hdr From:'
send-hook ~C'^b@b\.b$' my_hdr from: c@c.c

In Example 4.5, “Specifying a default hook”, by default the value of $from and $realname is not overridden. When sending messages either To: or Cc: to <b@b.b>, the From: header is changed to <c@c.c>.

6.1. Message Matching in Hooks

Hooks that act upon messages ( message-hook , reply-hook , send-hook , send2-hook , save-hook , fcc-hook ) are evaluated in a slightly different manner. For the other types of hooks, a regular expression is sufficient. But in dealing with messages a finer grain of control is needed for matching since for different purposes you want to match different criteria.

Mutt allows the use of the search pattern language for matching messages in hook commands. This works in exactly the same way as it would when limiting or searching the mailbox, except that you are restricted to those operators which match information Mutt extracts from the header of the message (i.e., from, to, cc, date, subject, etc.).

For example, if you wanted to set your return address based upon sending mail to a specific address, you could do something like:

send-hook '~t ^me@cs\.hmc\.edu$' 'my_hdr From: Mutt User <user@host>'

which would execute the given command when sending mail to .

However, it is not required that you write the pattern to match using the full searching language. You can still specify a simple regular expression like the other hooks, in which case Mutt will translate your pattern into the full language, using the translation specified by the $default_hook variable. The pattern is translated at the time the hook is declared, so the value of $default_hook that is in effect at that time will be used.

6.2. Mailbox Matching in Hooks

Hooks that match against mailboxes ( folder-hook , mbox-hook ) apply both regular expression syntax as well as mailbox shortcut expansion on the regexp parameter. There is some overlap between these, so special attention should be paid to the first character of the regexp.

# Here, ^ will expand to "the current mailbox" not "beginning of string":
folder-hook ^/home/user/Mail/bar "set sort=threads"
# If you want ^ to be interpreted as "beginning of string", one workaround
# is to enclose the regexp in parenthesis:
folder-hook (^/home/user/Mail/bar) "set sort=threads"
# This will expand to the default save folder for the alias "", which
# is probably not what you want:
folder-hook "set sort=threads"
# A workaround is to use parenthesis or a backslash:
folder-hook ( "set sort=threads"
folder-hook '\' "set sort=threads"

Keep in mind that mailbox shortcut expansion on the regexp parameter takes place when the hook is initially parsed, not when the hook is matching against a mailbox. When Mutt starts up and is reading the .muttrc, some mailbox shortcuts may not be usable. For example, the "current mailbox" shortcut, ^, will expand to an empty string because no mailbox has been opened yet. Mutt will issue an error for this case or if the mailbox shortcut results in an empty regexp.

7. Managing the Environment

You can alter the environment that Mutt passes on to its child processes using the setenv and unsetenv operators. (N.B. These follow Mutt-style syntax, not shell-style!) You can also query current environment values by prefixing a ? character.

setenv TERM vt100
setenv ORGANIZATION "The Mutt Development Team"
unsetenv DISPLAY
setenv ?LESS

8. External Address Queries

Mutt supports connecting to external directory databases such as LDAP, ph/qi, bbdb, or NIS through a wrapper script which connects to Mutt using a simple interface. Using the $query_command variable, you specify the wrapper command to use. For example:

set query_command = " %s"

The wrapper script should accept the query on the command-line. It should return a one line message, then each matching response on a single line, each line containing a tab separated address then name then some other optional information. On error, or if there are no matching addresses, return a non-zero exit code and a one line error message.

An example multiple response output:

Searching database ... 20 entries ... 3 matching:           Michael Elkins  mutt dude       Brandon Long    mutt and more        Thomas Roessler mutt pgp

There are two mechanisms for accessing the query function of Mutt. One is to do a query from the index menu using the <query>function (default: Q). This will prompt for a query, then bring up the query menu which will list the matching responses. From the query menu, you can select addresses to create aliases, or to mail. You can tag multiple addresses to mail, start a new query, or have a new query appended to the current responses.

The other mechanism for accessing the query function is for address completion, similar to the alias completion. In any prompt for address entry, you can use the <complete-query>function (default: ^T) to run a query based on the current address you have typed. Like aliases, Mutt will look for what you have typed back to the last space or comma. If there is a single response for that query, Mutt will expand the address in place. If there are multiple responses, Mutt will activate the query menu. At the query menu, you can select one or more addresses to be added to the prompt.

9. Mailbox Formats

Mutt supports reading and writing of four different local mailbox formats: mbox, MMDF, MH and Maildir. The mailbox type is auto detected, so there is no need to use a flag for different mailbox types. When creating new mailboxes, Mutt uses the default specified with the $mbox_type variable. A short description of the formats follows.

mbox . This is a widely used mailbox format for UNIX. All messages are stored in a single file. Each message has a line of the form:

From Fri, 11 Apr 1997 11:44:56 PST

to denote the start of a new message (this is often referred to as the From_ line). The mbox format requires mailbox locking, is prone to mailbox corruption with concurrently writing clients or misinterpreted From_ lines. Depending on the environment, new mail detection can be unreliable. Mbox folders are fast to open and easy to archive.

MMDF . This is a variant of the mbox format. Each message is surrounded by lines containing ^A^A^A^A(four times control-A's). The same problems as for mbox apply (also with finding the right message separator as four control-A's may appear in message bodies).

MH . A radical departure from mbox and MMDF , a mailbox consists of a directory and each message is stored in a separate file. The filename indicates the message number (however, this is may not correspond to the message number Mutt displays). Deleted messages are renamed with a comma ( ,) prepended to the filename. Mutt detects this type of mailbox by looking for either .mh_sequences or .xmhcache files (needed to distinguish normal directories from MH mailboxes). MH is more robust with concurrent clients writing the mailbox, but still may suffer from lost flags; message corruption is less likely to occur than with mbox/mmdf. It's usually slower to open compared to mbox/mmdf since many small files have to be read (Mutt provides Section 7.1, “Header Caching” to greatly speed this process up). Depending on the environment, MH is not very disk-space efficient.

Maildir . The newest of the mailbox formats, used by the Qmail MTA (a replacement for sendmail). Similar to MH , except that it adds three subdirectories of the mailbox: tmp , new and cur . Filenames for the messages are chosen in such a way they are unique, even when two programs are writing the mailbox over NFS, which means that no file locking is needed and corruption is very unlikely. Maildir maybe slower to open without caching in Mutt, it too is not very disk-space efficient depending on the environment. Since no additional files are used for metadata (which is embedded in the message filenames) and Maildir is locking-free, it's easy to sync across different machines using file-level synchronization tools.

10. Mailbox Shortcuts

There are a number of built in shortcuts which refer to specific mailboxes. These shortcuts can be used anywhere you are prompted for a file or mailbox path or in path-related configuration variables. Note that these only work at the beginning of a string.

Table 4.11. Mailbox shortcuts

Shortcut Refers to...
! your $spoolfile(incoming) mailbox
> your $mbox file
< your $record file
^ the current mailbox
-or !! the file you've last visited
~ your home directory
=or + your $folder directory
@alias to the default save folder as determined by the address of the alias

For example, to store a copy of outgoing messages in the folder they were composed in, a folder-hook can be used to set $record:

folder-hook . 'set record=^'

11. Handling Mailing Lists

Mutt has a few configuration options that make dealing with large amounts of mail easier. The first thing you must do is to let Mutt know what addresses you consider to be mailing lists (technically this does not have to be a mailing list, but that is what it is most often used for), and what lists you are subscribed to. This is accomplished through the use of the lists and subscribe commands in your .muttrc.

Now that Mutt knows what your mailing lists are, it can do several things, the first of which is the ability to show the name of a list through which you received a message (i.e., of a subscribed list) in the index menu display. This is useful to distinguish between personal and list mail in the same mailbox. In the $index_format variable, the expando %L will print the string To <list> when list appears in the To field, and Cc <list> when it appears in the Cc field (otherwise it prints the name of the author).

Often times the To and Cc fields in mailing list messages tend to get quite large. Most people do not bother to remove the author of the message they reply to from the list, resulting in two or more copies being sent to that person. The <list-reply>function, which by default is bound to L in the index menu and pager , helps reduce the clutter by only replying to the known mailing list addresses instead of all recipients (except as specified by Mail-Followup-To, see below).

Mutt also supports the Mail-Followup-To header. When you send a message to a list of recipients which includes one or several subscribed mailing lists, and if the $followup_to option is set, Mutt will generate a Mail-Followup-To header which contains all the recipients to whom you send this message, but not your address. This indicates that group-replies or list-replies (also known as followups) to this message should only be sent to the original recipients of the message, and not separately to you - you'll receive your copy through one of the mailing lists you are subscribed to.

Conversely, when group-replying or list-replying to a message which has a Mail-Followup-To header, Mutt will respect this header if the $honor_followup_to configuration variable is set. Using list-reply will in this case also make sure that the reply goes to the mailing list, even if it's not specified in the list of recipients in the Mail-Followup-To.


When header editing is enabled, you can create a Mail-Followup-To header manually. Mutt will only auto-generate this header if it doesn't exist when you send the message.

The other method some mailing list admins use is to generate a Reply-To field which points back to the mailing list address rather than the author of the message. This can create problems when trying to reply directly to the author in private, since most mail clients will automatically reply to the address given in the Reply-To field. Mutt uses the $reply_to variable to help decide which address to use. If set to ask-yes or ask-no , you will be prompted as to whether or not you would like to use the address given in the Reply-To field, or reply directly to the address given in the From field. When set to yes , the Reply-To field will be used when present.

You can change or delete the X-Label: field within Mutt using the edit-label command, bound to the y key by default. This works for tagged messages, too. While in the edit-label function, pressing the <complete> binding (TAB, by default) will perform completion against all labels currently in use.

Lastly, Mutt has the ability to sort the mailbox into threads. A thread is a group of messages which all relate to the same subject. This is usually organized into a tree-like structure where a message and all of its replies are represented graphically. If you've ever used a threaded news client, this is the same concept. It makes dealing with large volume mailing lists easier because you can easily delete uninteresting threads and quickly find topics of value.

12. Keyword Management

Mutt has supported textual labels (usually known as X-Labels after the header that we use to store them) for many years. Since we initially added support for X-Labels, however, the larger community has evolved more common ways of using and managing such labels, often known as keywords or tags.

If you are new to Mutt or to using keywords in Mutt, you only need to know that the <edit-label> binding will edit keywords, and that you can search for keywords using the ~y pattern, and use the %y expando to display it in your $index_format. You also can sort by keyword. Keywords that you set will be stored to the X-Label:header by default.

If you've been using X-Labels for a while, things have grown slightly. Mutt still supports X-Labels much as it has since 2000, but the scope of this support has expanded to support three additional header-based techniques for storing keyword metadata on messages:


Informal design; space-delimited keywords


Informal design used by Mozilla-based agents; space-delimited keywords


Standardized in RFC2822 (2001); comma-space-delimited keywords


Mutt-specific design; freeform text (but see $xlabel_delimiter)

With X-Label, mutt's only notion of a message keyword was the literal string value of the X-Label header. Under the new, integrated support, each message may have a list of distinct message keywords. When reading keywords from one of the headers in the list above, the header value is split on the indicated delimiter (space or comma-space) for X-Keywords:, X-Mozilla-Keys:, and Keywords:. By default, X-Label: is parsed as a single keyword. By setting $xlabel_delimiter, you can force splitting of X-Label: as well.

Two boolean variables control how keywords are saved when writing messages to a mailbox. The default settings preserve backward compatibility within mutt completely, but by changing these values you can transition to more standard keyword storage. $keywords_legacy, if set, will tell mutt to use only "legacy" headers -- i.e., X-Keywords:, X-Mozilla-Keys, Keywords, or X-Label:. Keywords will be saved to whichever header was in use by the message the keyword was read from. If $keywords_standard is set, keywords will be saved without exception to the standard Keywords:header. (If both are set, both will be used; if both are unset, legacy headers are used.) Additionally, $xlabel_delimiter is used to format the X-Label: header on saves.

To migrate completely to the new standard, unset $keywords_legacy and set $keywords_standard, and set $xlabel_delimiter either to what you currently use to delimit keywords in X-Labels, or to (comma space).

Note that it is common practice to insert X-Label:or other keyword headers from proxmail or other mail filters. This is a useful trick for categorizing messages en masse as they are delivered to your inbox, and it is fully compatible with the new keywords code.

13. Display Munging

Working within the confines of a console or terminal window, it is often useful to be able to modify certain information elements in a non-destructive way -- to change how they display, without changing the stored value of the information itself. This is especially so of message subjects, which may often be polluted with extraneous metadata that either is reproduced elsewhere, or is of secondary interest.

subjectrx pattern replacement
unsubjectrx{ * | pattern }

subjectrx specifies a regular expression pattern which, if detected in a message subject, causes the subject to be replaced with the replacement value. The replacement is subject to substitutions in the same way as for the spam command: %L for the text to the left of the match, %R for text to the right of the match, and %1 for the first subgroup in the match (etc). If you simply want to erase the match, set it to %L%R. Any number of subjectrx commands may coexist.

Note this well: the replacement value replaces the entire subject, not just the match!

unsubjectrx removes a given subjectrx from the substitution list. If *is used as the pattern, all substitutions will be removed.

Example 4.6. Subject Munging

# Erase [rt #12345] tags from Request Tracker (RT) e-mails
subjectrx '\[rt #[0-9]+\] *' '%L%R'
# Servicedesk is another RT that sends more complex subjects.

# Keep the ticket number.

subjectrx '\[servicedesk #([0-9]+)\] ([^.]+)\.([^.]+) - (new|open|pending|update) - ' '%L[#%1] %R'
# Strip out annoying [listname] prefixes in subjects
subjectrx '\[[^\]]*\]:? *' '%L%R'

14. New Mail Detection

Mutt supports setups with multiple folders, allowing all of them to be monitored for new mail (see Section 14, “Monitoring Incoming Mail” for details).

14.1. How New Mail Detection Works

For Mbox and Mmdf folders, new mail is detected by comparing access and/or modification times of files: Mutt assumes a folder has new mail if it wasn't accessed after it was last modified. Utilities like biff or frm or any other program which accesses the mailbox might cause Mutt to never detect new mail for that mailbox if they do not properly reset the access time. Other possible causes of Mutt not detecting new mail in these folders are backup tools (updating access times) or filesystems mounted without access time update support (for Linux systems, see the relatime option).


Contrary to older Mutt releases, it now maintains the new mail status of a folder by properly resetting the access time if the folder contains at least one message which is neither read, nor deleted, nor marked as old.

In cases where new mail detection for Mbox or Mmdf folders appears to be unreliable, the $check_mbox_size option can be used to make Mutt track and consult file sizes for new mail detection instead which won't work for size-neutral changes.

New mail for Maildir is assumed if there is one message in the new/subdirectory which is not marked deleted (see $maildir_trash). For MH folders, a mailbox is considered having new mail if there's at least one message in the unseen sequence as specified by $mh_seq_unseen. Optionally, $new_mail_command can be configured to execute an external program every time new mail is detected in the current inbox.

Mutt does not poll POP3 folders for new mail, it only periodically checks the currently opened folder (if it's a POP3 folder).

For IMAP, by default Mutt uses recent message counts provided by the server to detect new mail. If the $imap_idle option is set, it'll use the IMAP IDLE extension if advertised by the server.

The $mail_check_recent option changes whether Mutt will notify you of new mail in an already visited mailbox. When set (the default) it will only notify you of new mail received since the last time you opened the mailbox. When unset, Mutt will notify you of any new mail in the mailbox.

14.2. Polling For New Mail

When in the index menu and being idle (also see $timeout), Mutt periodically checks for new mail in all folders which have been configured via the mailboxes command. The interval depends on the folder type: for local/IMAP folders it consults $mail_check and $pop_checkinterval for POP folders.

Outside the index menu the directory browser supports checking for new mail using the <check-new>function which is unbound by default. Pressing TAB will bring up a menu showing the files specified by the mailboxes command, and indicate which contain new messages. Mutt will automatically enter this mode when invoked from the command line with the -y option.

For the pager, index and directory browser menus, Mutt contains the <buffy-list>function (bound to . by default) which will print a list of folders with new mail in the command line at the bottom of the screen.

For the index, by default Mutt displays the number of mailboxes with new mail in the status bar, please refer to the $status_format variable for details.

When changing folders, Mutt fills the prompt with the first folder from the mailboxes list containing new mail (if any), pressing <Space>will cycle through folders with new mail. The (by default unbound) function <next-unread-mailbox>in the index can be used to immediately open the next folder with unread mail (if any).

14.3. Calculating Mailbox Message Counts

If $mail_check_stats is set, Mutt will periodically calculate the unread, flagged, and total message counts for each mailbox watched by the mailboxes command. This calculation takes place at the same time as new mail polling, but is controlled by a separate timer: $mail_check_stats_interval.

The sidebar can display these message counts. See $sidebar_format.

15. Editing Threads

Mutt has the ability to dynamically restructure threads that are broken either by misconfigured software or bad behavior from some correspondents. This allows to clean your mailboxes from these annoyances which make it hard to follow a discussion.

15.1. Linking Threads

Some mailers tend to forget to correctly set the In-Reply-To: and References: headers when replying to a message. This results in broken discussions because Mutt has not enough information to guess the correct threading. You can fix this by tagging the reply, then moving to the parent message and using the <link-threads>function (bound to & by default). The reply will then be connected to this parent message.

You can also connect multiple children at once, tagging them and using the <tag-prefix>command ( ;) or the $auto_tag option.

15.2. Breaking Threads

On mailing lists, some people are in the bad habit of starting a new discussion by hitting reply to any message from the list and changing the subject to a totally unrelated one. You can fix such threads by using the <break-thread>function (bound by default to #), which will turn the subthread starting from the current message into a whole different thread.

16. Delivery Status Notification (DSN) Support

RFC1894 defines a set of MIME content types for relaying information about the status of electronic mail messages. These can be thought of as return receipts.

To support DSN, there are two variables. $dsn_notify is used to request receipts for different results (such as failed message, message delivered, etc.). $dsn_return requests how much of your message should be returned with the receipt (headers or full message).

When using $sendmail for mail delivery, you need to use either Berkeley sendmail 8.8.x (or greater) a MTA supporting DSN command line options compatible to Sendmail: The -N and -R options can be used by the mail client to make requests as to what type of status messages should be returned. Please consider your MTA documentation whether DSN is supported.

For SMTP delivery using $smtp_url, it depends on the capabilities announced by the server whether Mutt will attempt to request DSN or not.

17. Start a WWW Browser on URLs

If a message contains URLs, it is efficient to get a menu with all the URLs and start a WWW browser on one of them. This functionality is provided by the external urlview program which can be retrieved at the configuration commands:

macro index \cb |urlview\n
macro pager \cb |urlview\n

18. Miscellany

This section documents various features that fit nowhere else.

Address normalization

Mutt normalizes all e-mail addresses to the simplest form possible. If an address contains a realname, the form Joe User <> is used and the pure e-mail address without angle brackets otherwise, i.e. just .

This normalization affects all headers Mutt generates including aliases.

Initial folder selection

The folder Mutt opens at startup is determined as follows: the folder specified in the $MAIL environment variable if present. Otherwise, the value of $MAILDIR is taken into account. If that isn't present either, Mutt takes the user's mailbox in the mailspool as determined at compile-time (which may also reside in the home directory). The $spoolfile setting overrides this selection. Highest priority has the mailbox given with the -f command line option.