5 Letter Word Generator

What This Generator Actually Does

This is a constraint filter, not a puzzle solver and not a hint service. It takes a dictionary of five-letter words, removes everything that violates the constraints you type, and shows you what is left. That is the whole mechanism, and it is worth understanding because it explains both what the tool is good at and where it can mislead you.

The dictionary loads once and every search runs in your browser. Nothing you type is sent anywhere, there is no account, and there is no request to a server between typing a letter and seeing a count. That matters for two reasons: it is fast, and it means the tool works the same way whether you are on a phone with a weak connection or at a desk.

What it does not do is tell you which of the results is the answer. It has no knowledge of today's puzzle, no scoring model, and no ranking. A list of candidates is exactly that: a list. The judgement about which word to try next is still yours, and the rest of this page is about making that judgement well.

The Five Inputs, and What Each One Really Constrains

The tool has five kinds of input. They are not equally powerful, and knowing the difference is the fastest way to stop producing empty lists.

The practical consequence: two position boxes plus a start letter will usually produce a list you can read in one glance, while five contains-letters and no positions can still leave hundreds of words. If your results feel unmanageable, add a position. If they feel empty, remove a position before you remove anything else.

When to Use a Generator and When to Use a Solver

These are different tools with different jobs, and confusing them wastes time in both directions.

A generator works from constraints you already have. You know some letters, some positions, and some exclusions, and you want the set of words that satisfies all of it. It is a breadth tool: it shows you the whole candidate space so you can choose. It is what you want when you are stuck between several plausible words and need to see whether a word you are imagining actually exists.

A solver works from a board state and a scoring model. It does not just enumerate candidates; it ranks them, usually by how much information each guess would gain or by how likely each is to be the answer. It is a depth tool: it tells you which single word to play next, not which words are legal.

In practice the two alternate. Early in a puzzle, when you have almost no constraints, a solver's opener advice is more useful than a generator, because a generator with no filters just prints the dictionary. Late in a puzzle, when you have four constraints and a short list, a generator is more useful than a solver, because you can see the exact candidates and apply your own sense of which is more likely.

The failure mode is using a generator as if it were a solver: typing in a couple of letters, getting two hundred results, and picking one at random. That is not filtering, it is browsing, and it usually costs a guess. If the list is long, the constraints are not ready yet.

Starts-With and Ends-With Are Stronger Than They Look

Players tend to over-use the contains field and under-use the affix fields. That is backwards. A contains constraint tells you a letter is somewhere; an affix constraint tells you exactly where several letters are, all at once.

Consider what "ends with ING" really does. It fixes three of five positions and, more importantly, it fixes the word's grammatical category. The remaining candidates are not "words containing I, N and G"; they are present participles with a two-letter stem, which is a much smaller and much more predictable set. The same is true of -ED, -ER, -LY, -TION (too long here), -AL, -IC. A known suffix is a structural fact, not just a letter fact.

Starts-with has a similar effect at the front of the word, where English is rich in consonant clusters: ST-, STR-, SP-, SC-, SH-, CH-, TH-, WH-, BR-, CR-, DR-, TR-, FL-, GL-, PL-, SL-. If your first two letters are green, the start field is doing the work of two position boxes at once, and it will often reveal that the cluster is the only plausible reading of the board.

One caveat: a start or end constraint that is too long is the most common cause of a zero-result search. "Starts with STR" plus "ends with E" is already three fixed letters plus a fixed final letter, leaving one free position. If you also require two contains letters, the search is likely impossible. When a search returns nothing, shorten the affix first.

The Position Boxes, and the Mistake People Make

The position boxes are the most precise input and the most frequently misused. Each box enforces one letter at one index. That is all. It does not tell the tool that the letter is present elsewhere, and it does not interact with the contains field unless you make it.

The mistake is duplication. If you know the third letter is A, you can put A in box three or put A in the contains field — but if you do both, the contains field still only requires one A, so the result is the same. Where it goes wrong is the reverse case: you know the word has two As, one of them in position three. If you put A in box three and type A once in contains, the tool will accept words with a single A in position three, because one A satisfies both constraints. To require the second A, type AA in the contains field.

That is a subtle point, and it is the reason the contains field counts repeats. A word-game board does not tell you "this letter appears once"; it tells you a tile is green or yellow. When you have a green A and a yellow A, you are looking at two tiles, and the tool needs to be told about both.

The second mistake is using position boxes for yellow letters. A yellow letter is present but not where you played it. Putting it in a position box asserts something you do not know. Use contains for yellows and position boxes only for greens.

The Empty-Result Problem

Every constraint tool has the same failure mode: the user adds one filter too many and concludes the tool is broken. It is almost never broken. It is usually over-constrained, and the fix is to remove constraints in the right order.

Here is the order that resolves empty results fastest:

1. Check for a contradiction between fields. A letter in the excludes field that also appears in contains, starts-with, ends-with or a position box will always produce zero results. The status line warns you when it detects this, but it is worth knowing why.
2. Shorten the longest affix. A four-letter ends-with is a very tight constraint. Drop one letter and search again.
3. Remove the least certain position box. Position boxes are absolute, and a single wrong green assumption kills the search. If you are not sure, move that letter to contains instead.
4. Remove the excludes that came from a guess you are unsure about. One misread grey is enough to eliminate the answer.
5. Reduce the contains field, longest letter first. A rare letter such as Q, X, Z or J in contains is a very strong filter on its own.

There is also a legitimate zero-result outcome, and it is valuable: if you are confident in every constraint and the list is empty, one of your assumptions is wrong. That is a real deduction and often worth more than a list of candidates. Before you loosen a filter, ask which assumption you are least sure about, and test that one deliberately rather than removing constraints at random.

One more honest limitation: an empty list can also mean the word is missing from the dictionary. Our list is broad but not universal, and it is a word-game lexicon rather than a complete English dictionary. If you are certain a word exists and it never appears, that is the likely explanation.

Why the Five-Letter Pool Is Not the Scrabble Dictionary

People often assume every word list is the same list. They are not, and the differences matter when you are choosing a guess.

A puzzle answer pool — the set a daily five-letter game draws its secret from — is curated. It tends to contain common, familiar words, because a puzzle that answers with an obscure dialect term is not a good puzzle. It is deliberately narrow.

An acceptance list — the set of guesses a game will allow you to enter — is much larger. It has to be, because players type all sorts of things, and rejecting a real word feels unfair. This list includes plurals, inflections, and plenty of words that would never be an answer.

A Scrabble-style lexicon is larger again and follows a different philosophy: it includes obscure, archaic, dialect and loan words, and it excludes proper nouns and some hyphenated forms. It is built for competitive play, not for puzzle fairness. Words that are valid in one lexicon are routinely invalid in another.

Our generator uses a broad word-game list of five-letter entries. It is wider than a typical answer pool and narrower than a full competitive lexicon. The practical consequences:

• A result may not be an accepted answer. If a word looks obscure, it may be in the list but never chosen as a puzzle answer. Use the list as candidates, not as a promise.
• A valid word may be missing. Rare forms and very recent words are the usual gaps. If your list is empty and you are sure of a word, trust your knowledge over the tool.
• Plurals and inflections are included. That is useful for finding candidates and unhelpful if you assume every result is a plausible answer. A five-letter plural ending in S is usually a legitimate guess and rarely the solution.
• Obscure entries are not errors. The list is a lexicon. It contains words you have never seen, and that is a feature of word-game dictionaries rather than a bug in the filter.

If you want to check whether a word is likely to be an answer rather than merely a valid guess, the honest test is familiarity: would a general-audience puzzle use it? That judgement is not in the dictionary and never will be.

Reading a Result List Without Wasting a Guess

Once the list is short enough to read, the question changes from "which words fit" to "which word should I play". Those are different questions, and the answer is not always the first item alphabetically.

Prefer the word that splits the list. If the candidates are BAKE, CAKE, FAKE, LAKE, MAKE, RAKE, SAKE, TAKE and WAKE, any of them could be the answer, but only one guess can be played. Choosing the word whose first letter is most common is not the point; the point is that after the guess, the feedback should eliminate as many of the others as possible. In that set, almost any choice eliminates most of the list, because the words differ only in the first letter. In a set where the words differ in several positions, prefer a guess that tests the position with the most variation.

Prefer real words over probes when the list is short. A probe is a word chosen only for its letters. Probes are valuable early, when the list is huge, and wasteful late, when the list is small enough that a guess could simply be right. Once you are under ten candidates, every guess should be a candidate.

Watch for trap sets. Some candidate sets are designed to burn guesses: four words differing by one letter, where you can only test one per turn. In that situation, the correct play is often a word outside the set that tests the differing letter in the differing position while remaining a plausible guess. That is the one case where a non-candidate is justified late.

Check the count before you read the list. A list of eight is a decision. A list of four hundred is a signal that you need another constraint. The count is printed above the results for exactly this reason.

Three Worked Examples

Example one: one green, one yellow

You know the third letter is A, and the word contains an R somewhere else. Enter A in the third position box, R in contains, and generate. The list will be every five-letter word with A in the middle and an R anywhere else. Read it looking for words that also fit the pattern of your other feedback, and use the excludes field to remove letters you have already ruled out. This is the most common real-world use of the tool, and it usually produces a list short enough to scan.

Example two: a suffix and a ruled-out vowel

You believe the word ends in -LY and you know E is not in it. Enter LY in ends-with and E in excludes. The list is adverbs without an E. Many will be implausible as answers, which is useful: the plausible ones stand out immediately. If the list is empty, your suffix assumption is probably wrong, and that is a genuine deduction rather than a tool failure.

Example three: a start cluster and a repeated letter

You know the word starts with ST and contains two As. Enter ST in starts-with and AA in contains. The double letter in the contains field is doing real work here: without it, the tool would accept words with a single A. This is the case where the repeat-counting behaviour matters, and it is worth trying once so you understand what the field does.

How This Pairs With the Other Pages

The generator is one of three tools on the site, and each covers a different shape of problem.

The full five-letter word list is the browseable reference. It is organised by starting letter and is the right place when you want to read rather than filter — for example, when you are studying which words exist in a category, or checking whether a half-remembered word is real. It is a dictionary view, not a search tool.

The word finder is the general pattern tool. It handles other lengths, wildcard patterns, and anagram-style input, so it is the right choice when your problem is not specifically five letters. If you are working on a seven-letter puzzle or a Wordscapes level, start there rather than trying to force this page to do it.

This generator sits between them: five letters only, constraint-based, with position boxes and a live count. Use it when you have a partially solved five-letter board. Use the list page when you have no constraints and want to browse. Use the word finder when the length is not five.

Copying the List, and a Sensible Workflow

The copy button puts every matching word on your clipboard, one per line, in alphabetical order. It copies the full result set, not just the 500 shown, so a large search does not silently lose words. If the clipboard is blocked by your browser, the status line says so and you can select the visible list by hand.

A workflow that works well on paper or in a notes app:

1. Generate with the constraints you are sure of. Copy the list.
2. Paste it into a scratch document and delete the implausible words. This is the step that uses your knowledge rather than the dictionary. Most of the value is here.
3. Mark the words that would split the remaining set if wrong. Those are your next guesses.
4. When the board changes, regenerate rather than editing the old list by hand. Manual editing is where stale candidates creep in.

If you are studying rather than solving, keep the list and write next to each word why it fits. That turns a filtered list into a record of your constraints, which is far more useful the next time you meet a similar pattern.

What the Tool Will Not Do

It will not rank candidates by likelihood, because likelihood depends on the puzzle's answer pool, which it does not know. It will not tell you today's answer. It will not score a guess or model future feedback. It will not handle words of other lengths, and it will not accept apostrophes, hyphens or spaces, because five-letter puzzle words do not contain them.

It also will not fix a wrong constraint. If you type a letter into a position box that is not actually green, the tool will faithfully return words that fit a board that does not exist. The output is only as good as the input, and the most common cause of a bad list is a misremembered tile rather than a dictionary gap.

Used within those limits, it does one thing well: it turns a partial board into a readable set of candidates in less time than it takes to think of them one by one. That is the entire promise, and it is enough.