What Clueword Is, and Why It Is Harder Than Wordle
Clueword is Wordle with the loop reversed. In Wordle you type a word, the board colours it, and you use that feedback to shape the next guess. In Clueword the coloured rows are already printed when the puzzle loads. One to five rows of five real letters, each tile coloured against a hidden answer that you never see. Nothing is hidden about the constraints — everything is hidden about the word.
Your job is to find any five-letter word that satisfies all of the rows simultaneously. That is a deduction puzzle, not a guessing game, and the difference changes how you should spend your time. In Wordle a wrong guess is a purchase: you pay one of six attempts and you buy information. In Clueword a wrong attempt buys almost nothing, because the information was already on the board. You either read the rows correctly or you do not.
The daily puzzle uses three clue rows and four tries. The levels ladder trades information against allowance in the opposite direction to most games: low levels give you one or two rows and four tries, the middle of the ladder gives you three rows and three tries, and the top gives you four or five rows with only two tries. More clues, fewer mistakes permitted. On a five-row board with two tries, there is no room to test a hunch.
Two consequences follow, and they are worth holding on to:
- There is no single intended answer. The game validates your word against every printed row rather than against the word its generator started from. If your word reproduces all the colours, you win, even if it is not the word the puzzle was built around. The generator's own word is only disclosed in the end-of-puzzle message, once you have already won or run out of tries.
- There is no reveal-letter control. Some word games let you spend a resource to expose a letter. Clueword does not. The clue rows are the entire information set, which is exactly why a solver is useful here: it does not replace the clues, it enumerates everything the clues still allow.
Why does that make Clueword harder than Wordle for most players? Because Wordle is forgiving of a bad first guess and Clueword is not forgiving of a bad reading. A Wordle player can recover from a weak opener in three moves. A Clueword player who misreads one grey tile can eliminate the actual answer before they ever consider it, and no amount of vocabulary will bring it back. The skill being tested is closer to sudoku than to Scrabble: read the givens precisely, then propagate.
Vocabulary still matters, but later in the process. Once the constraints have narrowed the board to a handful of letter patterns, you have to recognise which real words fit those patterns. That is the moment when a wide word list helps, and it is also the moment where this page does its work.
How the Coloured Clue Rows Work
Every clue row is itself a real five-letter word. The game scores that word against the hidden answer using exactly the same rules Wordle uses for a guess, and prints the result. So a clue row is not a cryptic clue and it is not a definition — it is a completed guess with its feedback attached.
| Colour | What it says about that cell | What it does not say |
|---|---|---|
| Green | The answer has this letter in this exact cell. The letter is present and the cell is closed to every other letter. | Nothing about the other cells. A green is one fact, not a fact about the whole row. |
| Yellow | The answer contains this letter somewhere else. Two facts at once: the letter is required, and this cell is banned for it. | It does not say how many copies the answer holds, only that at least one of this row's copies was matched here. |
| Grey | This particular copy of the letter found no match in the answer. It also rules the letter out of this exact cell, because an exact match would have been scored green. | It does not automatically ban the letter. That depends on whether another copy of the same letter in the same row is green or yellow. |
That last row of the table is where most of the difficulty lives, and it is the rule this page implements most carefully. A grey tile is a statement about one copy of a letter, not necessarily about the letter as a whole.
There is one more positional fact worth spelling out, because it is easy to miss and it is genuinely useful. If a row shows a grey tile in cell three, the answer cannot have that letter in cell three. The scoring pass checks exact matches first, so any cell that matches exactly is green by definition. A grey tile therefore carries two pieces of information: a count limit, and a positional ban. Solvers that only count letters throw the positional ban away and search a wider space than they need to.
The Grey Duplicate Rule, With Worked Examples
Here is the rule in one sentence: a grey tile caps how many copies of that letter the answer may contain; it only bans the letter outright when no copy of that letter is green or yellow anywhere in the same row.
Three examples make the difference concrete.
Example one — grey does not ban a letter that is also green. Suppose a row reads SPEED and the answer is SPEND. The tiles score green, green, green, grey, green: S, P and E match exactly, the second E finds nothing left to match, and D matches exactly. The grey E does not mean “E is banned”. It means the answer has exactly one E, and that E is already pinned in cell three. A solver that bans E on sight of that grey tile would fail a board it should solve instantly.
Example two — grey does not ban a letter that is also yellow. In the sample this page can load, the row BERRY scores grey, grey, yellow, grey, grey against the answer POLAR. The first R is yellow because the answer has an R; the second R is grey because there is no second R left to match. Read together, the two R tiles say “exactly one R, and it is not in cell three”.
Example three — a grey duplicate sets an exact count. A row reading LULLS against the answer LEGAL scores green, grey, yellow, grey, grey. Three L tiles: one green in cell one, one yellow in cell three, one grey in cell four. Because a grey copy exists, the answer contains exactly two Ls — the green plus the yellow — and the third copy is surplus. If the same row had no grey L at all, the correct reading would be “at least two Ls”, not “exactly two”.
Notice what the three examples have in common: the grey tile is always true, and the naive interpretation of it is always too strong. The count interpretation is the one that survives across rows.
How the solver turns that into arithmetic
For every letter, the page computes two numbers from the rows you entered.
- A lower bound. For each row, count how many copies of the letter are green or yellow. The answer needs at least that many. Across rows, take the maximum, because every row has to be satisfied at once.
- An upper bound, when a grey copy exists. In a row that contains a grey copy of the letter, the number of green plus yellow copies is an exact count, not a minimum. Across such rows, take the minimum, because the tightest cap wins.
A letter with no grey copy anywhere has no upper bound from the board, so only the lower bound applies. A letter that is grey everywhere and never green or yellow has an upper bound of zero, which is the same as a ban. If at any point the upper bound falls below the lower bound, the rows contradict each other and no word can satisfy them — the page says so rather than inventing a near miss.
The rule generalises to something worth memorising: greens and yellows raise the count, greys cap it, and the tightest cap wins. Everything else on a Clueword board is bookkeeping on top of those three moves.
Reading a Full Five-Row Clue Set
The sample button loads this exact set, which is a real five-row board with two surviving answers. Reading it by hand is the best way to see how the rows interact.
| Row | Colours | What it contributes |
|---|---|---|
BERRY | grey, grey, yellow, grey, grey | One R, not in cell three. B, E and Y are dead. The second R is grey because the first already used the answer's only R. |
ROVER | grey, green, grey, grey, green | O fixed in cell two, R fixed in cell five. That green R is what stops the grey Rs from banning R entirely. |
JERRY | grey, grey, yellow, grey, grey | Repeats the R count: exactly one R, still not in cell three. J is dead. |
BURST | grey, grey, yellow, grey, grey | R again, same count. U, S and T are dead. |
DAILY | grey, yellow, grey, yellow, grey | A is required and not in cell two; L is required and not in cell four. D and I are dead. |
Now propagate, in the order that costs least.
Step one: write the frame. Two greens, so cells two and five are settled: ? O ? ? R. No other row supplies a green, which is typical of a five-row board — the information is spread across counts and bans rather than concentrated in fixed positions.
Step two: collect true bans. B, E, Y, V, J, U, S, T, D and I are grey everywhere they appear, with no green or yellow copy anywhere. Ten letters gone, and the board is already much narrower than it looks.
Step three: handle the duplicate. R is grey in three rows, which is exactly the trap. But R is green in cell five of ROVER, so the greys cap the count at one and ban the letter from cells one, three and four. R survives, pinned at the end. A player who crosses out R here has just deleted both answers.
Step four: place the yellows. A must appear and cannot sit in cell two. L must appear and cannot sit in cell four. Cells two and five are taken, so A and L have to share cells one, three and four between them, with one spare cell for a fifth letter.
Step five: intersect with real words. The frame is ? O ? ? R with A and L inside it and ten letters unavailable. Try the plausible placements and the board collapses quickly: putting L in cell three and A in cell four gives ? O L A R, and once the banned letters are removed the only survivors of that shape are MOLAR and POLAR. Putting L in cell one gives L O ? ? R, which leaves nothing once A is required and the bans are applied. That is the whole puzzle — and both survivors fit every row, so both would be accepted by the game.
This is the shape of a good Clueword solve: the first four steps are mechanical and take under a minute, and the fifth step is where vocabulary enters. The solver on this page automates steps one to four and then searches a 12,478-word list for step five, which is why it can return an answer you would never have thought of and still be right.
What This Solver Does, and What It Cannot Do
The solver reads the rows you enter, derives the constraint set described above, filters a dictionary of 12,478 five-letter words, and then re-checks every survivor by replaying the game's own scoring function against all of your rows. The re-check matters: the constraints are a fast filter, but the scoring function is the authority, and a candidate is only shown if it reproduces every colour you entered, cell by cell.
It reports the total number of matches, not just the ones it prints. If forty words fit and the page shows the first forty, the count says forty. If four hundred fit, the count says four hundred and the list is capped, so a long list is never disguised as a short one.
It also explains its own reasoning: which cells are fixed, which letters are required, which letters are capped, and — when nothing fits — which clues disagree. That last part is the honest core of the tool. A solver that always returns something is worse than useless on a misread board, because it lets you spend a try on a word the board has already ruled out.
What it cannot do:
- It cannot read your screen. You type the rows; a mistyped letter or a colour set one cell to the left produces a confident, wrong answer. If a result looks bizarre, re-check the row before doubting the logic.
- It cannot rank candidates by likelihood. There is no intended answer to rank toward. Every word it returns reproduces your rows, and the game accepts any of them. Treating the first entry as “the” answer is a misunderstanding of the game, not a shortcut.
- It does not track your tries. It has no idea whether you have four attempts left or one. That is deliberate: the page is a deduction aid, not a game client.
- It cannot fix an impossible pattern. Some colour arrangements cannot occur at all — a grey copy sitting to the left of a yellow copy of the same letter, for instance, because scoring colours the leftmost unmatched copy first. The page detects that and says so instead of returning words that would not actually reproduce the row.
- It uses a wide dictionary. The list includes obscure entries, so some results will look like typos. They are legal five-letter words in the same list the game validates against, which means the game will accept them even when the puzzle designer never considered them.
One boundary is worth stating plainly because it surprises people. A board with one clue row can easily have hundreds of valid answers, and the solver will say so. That is not the tool failing; it is the board being genuinely underdetermined. Difficulty in Clueword comes from stacking rows until the intersection is small, and a single row is barely a constraint at all.
How to Solve Clueword Without a Solver
The tool is a convenience. The method underneath it is short enough to run in your head, and it is faster than typing once you have used it a few times.
1. Read every row before writing anything. The common failure is to start with the first row, build a theory, and then force the later rows to agree with it. Read all the rows once, and note the greens across the whole board first. Greens are the only clues that fix a cell, so they define the skeleton everything else hangs on.
2. Start with the most constrained row. Not the first row, and not the longest word — the row with the most non-grey tiles, or the row whose non-grey letters are the most distinctive. A row of five greens hands you the answer. A row with four yellows tells you the answer is an anagram of those letters. A row that is almost entirely grey tells you very little beyond a handful of bans, even though it looks dramatic.
3. Turn greys into bans, carefully. For each grey tile, ask one question: is this letter green or yellow anywhere in the same row? If no, delete the letter from your alphabet. If yes, write down the count instead — the answer holds exactly as many copies as that row matched, and the grey copy is surplus.
4. Count duplicates explicitly. Keep a tally per letter rather than a yes/no list. “Contains L” and “contains exactly one L” and “contains at least two Ls” are three different constraints, and the difference decides boards. When two rows disagree about a count, the board is either misread or invalid; there is no third option.
5. Place the yellows into what is left. A yellow is a requirement plus a ban, so each one removes a cell from consideration. After the greens, the greys and the bans, a yellow often has exactly one legal cell left. When two yellows compete for the same cell, one of them is misread.
6. Check the finished word against every row before you submit. Not against the rows you used to build it — against all of them, cell by cell, including the counts. This is the step players skip when they are close, and it is the step that costs the try. With four attempts on the daily and two at the top of the ladder, a verification pass is always cheaper than a wrong submission.
If logic stalls, do not guess to “see what happens”. Feed the frame you have into a word list instead. That is exactly what the tool above does, and doing it by hand with a pattern such as ? O L A R is the same operation at a smaller scale. The word finder handles the pattern side if you want to keep the deduction manual.
Common Mistakes That Break a Board
These are the errors that actually cost attempts, roughly in order of how often they appear.
Treating grey as an absolute ban. This is the big one, and the sample board on this page is built to punish it. R is grey in three of the five rows. Ban it and the frame becomes impossible, because R is green in cell five and both surviving answers end in R. The fix is one question per grey tile: does the same letter appear green or yellow in this row?
Ignoring a later yellow because an earlier row seemed to settle the letter. Rows are simultaneous constraints, not a sequence of hints. If row one suggests a letter is absent and row four shows it yellow, row four wins — and if the two cannot be reconciled, one of your readings is wrong. Later rows do not override earlier ones; all of them apply at once.
Missing duplicates entirely. A letter that appears twice in a row is two constraints, not one. Two yellows of the same letter mean the answer holds at least two copies. A yellow plus a grey means exactly one. Reading the row as “contains this letter” throws away the count, which is often the only thing that separates the final two candidates.
Reading cell positions off by one. When you translate a row into a pattern, count the cells. A green in cell four written down as cell three quietly bans the right letter from the right place and admits it in the wrong one. On the sample, O is in cell two and R is in cell five; shifting either by one destroys the frame.
Building a word from one row and checking the others afterwards. This is how confirmation bias enters. You find a word that fits the row you were staring at, then bend your reading of the remaining rows to keep it. Build from the intersection, and check against every row at the end.
Assuming the generator's word is the only answer. It is not. The game accepts any word in its list that reproduces the colours, and the generator's word is only disclosed in the end-of-puzzle message. If a solver gives you a word you did not expect and it fits every row, the board agrees with you and the puzzle is solved.
Spending a try to test a letter the board already answers. On a two-try board this is fatal. If a letter is green anywhere, you know it is in the word; if it is grey with no green or yellow twin, you know it is out. A test guess should only happen when the constraints genuinely leave several words and you cannot check them against the board — which is rare once the rows are read properly.
Clueword Solver FAQ
What does a Clueword solver do?
It takes the coloured clue rows from a Clueword board and returns every five-letter word in its dictionary that would have produced all of those colours at once. It does not guess a hidden answer, because Clueword accepts any word that fits every row.
Why is a grey letter sometimes not banned?
A grey tile only means that one copy of the letter found no match. If the same letter is green or yellow elsewhere in the same row, the grey copy caps how many copies the answer holds instead of banning the letter. Banning it is the most common way to lose a Clueword board.
Why did the solver return no candidates?
Either the clues contradict each other or no word in the 12,478-word list reproduces them. The page names the rows that disagree when it can. The usual cause is a colour set on the wrong cell or a letter typed into the wrong row, so re-check the board before assuming the puzzle is broken.
How many answers does a Clueword board have?
It depends on the number of clue rows. One row can leave hundreds of valid words, three rows usually leave a handful, and five rows often leave one or two. The solver always reports the true total, even when it only prints part of the list.
Does the solver know today's Clueword answer?
No. It has no connection to the daily puzzle and no stored answers. It works only from the rows you type, which is why it can help on any board, including levels and archived puzzles, and why it cannot tell you the generator's original word unless your rows happen to narrow to it.
Related Clueword and Word Tools
If you want to play first, open the Clueword game. The Clueword rules guide covers the modes, the tries ladder and the full five-step method, and today's Clueword hints give graded nudges before the answer. For other boards, the Chainword solver and the codeword solver follow the same pattern of turning a printed grid into a candidate list, and the five-letter word lists browse the same vocabulary this page searches.