Word tool

Word Unscrambler & Exact Anagram Solver

Use the full rack or a subset. Add ? for a blank. Results are generated locally and paired with a complete tutorial below.

Functional browser tool

Build words from your letters

Enter 2–15 English letters; spaces are ignored. ? or * supplies one blank. Exact mode uses every tile and overrides the length range. Matching runs on your device using the ENABLE word list.

Examples: AELRST subsets · CART? exact · LET? exact

Try AELRST to find LATER, STALE, ALERT, STAR, RATE, and more.

How to use the solver well

The tool is most useful when you supply a clear constraint and then verify unfamiliar results.

Step 1

Enter the rack

Type letters only. Keep a blank as ? or *. Example: CART? contains four fixed letters and one flexible tile.

Step 2

Set the length

Choose Exact anagram to use every tile. For shorter words, leave it off and choose a length range.

Step 3

Check the context

Confirm rare results in the dictionary required by your game, class, or publication.

Worked example: AELRST

Sort the rack as A-E-L-R-S-T. A subset search finds three-letter foundations such as ART, RAT, SEA, and SET. Four-letter extensions include RATE, STAR, and TALE. Five-letter candidates include ALERT, ALTER, LATER, STALE, STEAL, SLATE, and TALES. Each candidate is accepted only when its letter counts fit the rack.

Exact mode: Exact mode keeps only six-letter candidates using every supplied letter. Subset mode returns shorter buildable words as well.

A reproducible routine: rack to best word in 7 steps

This routine works for any rack and takes about a minute. The running example is the rack ADEGRN — every count below was produced by this site's bundled ENABLE list (168,551 unique words, 2–15 letters), so you can check each number against the tool.

  1. Sort your rack and count duplicates. Write the letters in order: A-D-E-G-N-R. No letter repeats here, so each candidate may use every letter at most once.
  2. Choose the mode before you type. Tick Exact anagram only if you must play every tile. Otherwise leave it off — for this rack you will see 74 candidates instead of 4.
  3. Enter the rack and set a minimum length. Type ADEGRN with minimum length 2. For racks of 8 or more letters, raise the minimum to 4 or 5; the two- and three-letter words add noise without helping your decision.
  4. Read the longest group first. For ADEGRN the six-letter group has 4 words: DANGER, GANDER, GARDEN, RANGED. In a game these clear the most tiles; in study they form one anagram family worth learning as a set.
  5. Work downward and watch what you leave behind. The five-letter group has 7 words (ANGER, GRADE, GRAND, RAGED, RANGE, REDAN, REGNA). It shows what each choice costs: sometimes a shorter word that leaves a better tile mix is the stronger play.
  6. Look up every unfamiliar word. REDAN and REGNA may be new to you. A word you cannot define is a word a challenge can take away in stricter company — check its meaning before you play it.
  7. Verify against your game's list. This tool matches ENABLE. If your game uses a different official list, confirm the chosen word there. Presence here proves only that ENABLE contains the word; absence here proves nothing about validity.

What a real rack returns: ADEGRN by the numbers

The table below is not an estimate. It is the actual match total from running the rack ADEGRN through this site's ENABLE list in subset mode, minimum length 2:

Subset matches for the rack ADEGRN (168,551-word ENABLE list)
Word lengthMatchesExamples
2 letters12AD, AE, AG, AN, AR, DE, ED, EN
3 letters25AGE, AND, ANE, ARE, DAG, DEN, EAR, END
4 letters26AGED, AGER, DANG, DARE, DARN, DEAN, DEAR, DRAG
5 letters7ANGER, GRADE, GRAND, RAGED, RANGE, REDAN, REGNA
6 letters4DANGER, GANDER, GARDEN, RANGED
Total74

Two things stand out. First, the four-letter group is the largest — most racks hide their bulk in the middle lengths, so skimming only the longest group misses most of what a rack contains. Second, exact mode would return only the 4 six-letter words. When a search appears to find "nothing," the first thing to check is whether exact mode is still ticked from an earlier search.

Pitfalls we hit while testing this tool

Exact mode left on from a previous search. In testing, the most common "the tool found nothing" report came from users who had ticked Exact anagram earlier and forgotten it. Exact mode on a 15-letter rack almost always returns an empty list, and an empty list looks exactly like a broken tool. Treat an empty exact-mode result as a prompt to untick the box and search again before concluding anything is missing.

Long racks with minimum length 2 flood the list. A 15-letter rack in subset mode with the default minimum returns thousands of two- and three-letter words. Every one of them is a correct match — but the list becomes unusable. For racks of 8 or more letters, set the minimum length to 5; the results collapse to the words you would actually consider.

Pasted text brings invisible characters. We found that pasting a rack from a chat app or a screenshot OCR sometimes carried smart quotes or zero-width spaces. The tool shows a visible error instead of silently truncating the input, which is the intended behavior — the fix is on your side: retype the letters by hand. If you see the error, assume a stray character first, not a bug.

Underscores are not blanks. Several testers typed an underscore for a blank tile out of habit from other solvers. This tool accepts ? or * only; an underscore triggers the visible input error. Note that a blank-supplied letter scores zero and is highlighted in the results, so you can always see which letter the blank stood in for.

Algorithm in plain English

The solver counts every letter in the rack and every letter in a candidate. If a candidate asks for more copies than the rack contains, it is rejected unless a blank covers the difference. It then sorts accepted words by length, estimated tile value, and alphabetic order.

Limits and dictionary scope

This version uses ENABLE entries of 2–15 lowercase English letters, including uncommon words. It is not a current official tournament dictionary or a definition service. Read the source and limitations; verify unfamiliar entries in your required word list.

Common questions

How to read the result list

Results are grouped by length because length is what a rack rewards: a seven-letter play uses every tile, scores the tile sum plus any board premiums, and empties the rack for a fresh draw. Within a length group, look first for words built from letters you are glad to spend — duplicated letters, awkward pairings like two I's or a lone Q without its U. A shorter word that spends your worst tiles is often the better play, and it is nearly always the better vocabulary find, because the letters it leaves behind combine with more of the dictionary. If a result is new to you, check its definition before you play it; a word you can define is a word you will remember next game, and a word you cannot define is one a challenge will take away in stricter company.

From solver results to a real vocabulary

A solver session produces more words than anyone can memorise at once, so treat the list as raw material, not homework. Pick the three results that surprised you and write each one into a sentence of your own — the act of using a word once does more for retention than rereading it ten times. Notice families as you go: ALERT, ALTER and LATER are one anagram set, and learning them as a set teaches you the letter group AELRT, which recurs constantly in English racks. Over a few weeks of this, the solver stops being an answer machine and becomes a pattern trainer: you start seeing sets in your rack before you type anything, which is the actual skill word games reward.

Fair use in games and classrooms

Whether using a solver is cheating depends entirely on the agreement at the table. In a rated game, it is cheating, full stop. In casual play, it is whatever the players agreed before the first tile was drawn — many friendly games allow a solver for one "lifeline" play or for checking after a move. In study and in classrooms there is no controversy at all: working backwards from a rack to the words it hides is a legitimate spelling and vocabulary exercise, and teachers use exactly this method with letter tiles on a desk. This tool is published as a study and reference aid; what a fair game is, your group decides — decide it before you start, and the solver will never cause an argument.

Does the tool send my letters anywhere?

No. Matching happens in your browser.

Why is a word missing?

It may be outside ENABLE or excluded by your mode or length settings. The absence does not prove the word is invalid.

Does a blank score points?

Blank-supplied letters are highlighted and score zero. CAT from CA? has base score 4. Board multipliers, crossing words and bonuses are not included.

I ticked Exact anagram and got no results. Is the tool broken?

Almost certainly not. Exact mode keeps only words that use every tile, and most long racks have no full-length anagram at all. Untick Exact anagram and search again — if the subset list is full, the data is fine and the rack simply has no bingo.

Why are there thousands of results for a long rack?

Subset mode returns every buildable word at every allowed length, and a 12–15 letter rack genuinely contains thousands of them. Raise the minimum length to 5 or 6 to cut the list to the words you would actually consider.

Can I trust a result for tournament play?

No — not on this tool alone. The list is ENABLE, a broad historical lexicon, not the current official list of any tournament or app. Confirm the word in the list your game requires before you play it; a challenge can remove an unverified play.