Welcome to the Mystery Hunter's Word Search!

The Mystery Hunter's Word Search is a set of tools that allows you to search a number of dictionaries in a number of different ways.

Regular Expression Tool

The regular expression tool allows you to use extended regular expressions to find matches in the dictionary. A full regular expression tutorial is available here, but the cheatsheet included with the tool should be enough to get you started. The regexp tool is case-insensitive, and will max out at 1000 results.

Things to try:
^...aly...$ will find all 9-letter words with "aly" in the middle
^(...).*\1$ will find all words that start and end with the same 3 letters
(?=.*c.*)(?=.*r.*)(?=.*u.*)(?=.*d.*)(?=^c...$) will find all anagrams of "crud" that start with "c"

Anagram Tool

The anagram tool will generate multiple-word anagrams from a given set of letters. Unlike regexp anagrams (see the example above), this tool does not accept any wildcards; however, it includes the function to filter results with a regular expression.

It is highly recommended NOT to use large dictionaries when using this tool. OneLook, Compound and UK Cryptic dictionaries, in particular, are useless, because of how they are formatted (see Dictionaries below).

Things to try:
motherinlaw, specify start word as "woman" and maximum number of words as 2, and use the "Moby single words" dictionary.

Cryptogram Tool

The cryptogram tool will try to solve multiple-word cryptograms, with or without spaces. It uses an optimized dictionary, so the choice of a dictionary is not available in this tool. If you know some of the letter mappings, you can provide them for the tool in the supplied field. This tool is case-insensitive.

Things to try:
V unmgrkooc ed znckw kzorkt qcoddoy ior, check "trust the spaces". wlflgqoutozotmutbopbnoyulzotvpmsopyilhonstmd, without trusting the spaces.

Letterbanks Tool

This tool allows you to create multiple letterbanks and use them to construct word queries. Think of this as multiple interleaved anagrams with wildcards. In addition, you can combine letterbank references with regular expressions for more powerful searches.

Let's look at an example. Let's say you are working on a puzzle, and you know your answer should be 9 letters long, with the first letter being "c", the next three letters belonging to a group "aactmgh" and the last five letters belonging to the group "lyss." (four known and one unknown). You create two letterbanks: /1 = aactmgh, and /2 = lyss. -- and then build your query: (c)/1/1/1/2/2/2/2/2. Your result? Catalyses, catalysts or catalysis.

You can include regular expressions along with the letterbank references, but make sure that each regular expression is enclosed in parentheses. Keep in mind that something like (.*)/1/1/1(.*) is equivalent to (.*)/1/1/1, since the first regular expression group is greedy and will consume the word up until the last three characters (which are restricted to be checked against the letterbanks).

Things to try:
/1/1/1/2/2/2/2, with /1 = ret, and /2 = anis.

About Our Dictionaries

Most of the dictionaries we have available to search are in a one-word-per-line format, and their content is self-explanatory from their name. There are a few notable exceptions:

Compound Words and Phrases - this dictionary contains only multiple-word phrases, or words with dashes. Every line in the dictionary is formatted as follows (\t is regexp notation for Tab character):

somephrase\tsome phrase\t(4,6) or someword\tsome-word\t(4-4)

OneLook Word List - this is a mother of all word lists, graciously provided by the owner of onelook.com. This word list has 7.2 million entries, words and phrases, grouped and tagged by the number of dictionaries they appear in (Onelook indexes hundreds of dictionaries). Each line in the dictionary is formatted as follows:

word 101 or some_phrase 101

101 in either case is the number of dictionaries that index the word or phrase. Also note that the spaces in phrases are replaced with underscores.

IMPORTANT: Keep in mind that OneLook dictionary is humongous (it's a 140 Mb file -- our next largest dictionary is about 6.2 Mb), and therefore should be used sparingly, as searching it puts a heavy load on the server.