Wordle solver — filter candidate words from your clues
Enter each guess, then mark every letter green, yellow or grey. The list below updates as you go, showing every five-letter word that's still consistent with everything you've entered, plus a suggested next guess.
How the filtering works
Each coloured tile is a small piece of evidence about the answer, and the solver combines all of them, across every row, before deciding which words survive.
Green — this position is fixed
A green tile means the answer has that exact letter at that exact position. Every remaining candidate must match it there.
Yellow — present, but not here
A yellow tile means the letter is in the answer somewhere, just not at this position. It rules out one specific placement while requiring the letter to show up elsewhere in the word.
Grey — usually absent, but read the count, not just the colour
A plain grey tile for a letter that appears nowhere else in your guess means the answer has none of that letter. That's the easy case. The subtle case is a repeated letter: if you guess a word with the same letter twice and only one of the two tiles comes back green or yellow, the other one coming back grey does not mean the answer has zero of that letter. It means the answer has exactly as many as you've already confirmed — often one, sometimes two if both non-grey tiles are hits.
Worked example. Guess a word with a double L, and say the answer only has one L. One L tile comes back green (or yellow), confirming an L exists. The other L tile comes back grey — not because there's no L, but because there's no second L. A word with a single L, in the right places, still matches; a word with a double L is correctly ruled out. This solver counts letters per row, rather than treating grey as a blanket ban, so this case works the way the game actually scores it.
Every state is also marked with its own small glyph in the corner of the tile — a check for green, a dot for yellow, an × for grey — and announced in full to screen readers, so the colours are never the only way to read a tile's state.
Combining rows
Constraints from every row you've entered are merged: the tightest position requirements, the fullest set of required letters, and the lowest known cap on any letter that's come back grey after being confirmed. Adding another guess row only ever narrows the list further, never widens it.
A helper, not an answer key
This tool doesn't know today's answer and isn't connected to any specific game — it only knows the rules of green/yellow/grey and a general English word list. Used well, that makes it a decent second opinion rather than a shortcut: type in the clues you already have, see which of your own guesses are still logically possible, and use the suggestion as one reasonable option among several rather than the one true next move.
A common way to use it without spoiling the fun: keep playing normally, and only open this if you're genuinely stuck on a specific word, or if you want to double-check that a guess you're considering is actually still valid given the clues so far.
Picking a starter word
Before any clues exist, the best opening guess is a word made of five different, common letters — no repeats, so every tile teaches you something. Letters such as E, A, R, S, T, O, N, I and L turn up in a large share of five-letter English words, so a guess built mostly from those tends to split the candidate list well on the very first try. There's no single correct starter; once you have real clues, the live suggestion above is generally a stronger move than sticking with a fixed opener.
The general-dictionary caveat
This solver filters a general-purpose English word list of roughly 8,500 five-letter words, not the specific, curated answer list any one game actually draws from. That means two things worth knowing: it can show you a perfectly valid English word as a "candidate" that a particular game would never use as its solution, and it has no way to know which candidate is today's actual answer. What it can tell you, reliably, is which words are still logically consistent with the clues you've entered — that's a filtering job, not a lookup of a hidden answer.
Frequently asked questions
Does grey always mean the letter isn't in the word at all?
Not when the letter appears more than once in your guess. A grey tile means “no more copies of this letter beyond what’s already confirmed” — it caps the count, it doesn’t always zero it out. If you guess a word with a double letter and only one comes back green or yellow, the other coming back grey means the answer has exactly that many, not zero. This solver tracks the count per letter, not just presence or absence, so duplicate letters are handled correctly.
Is using this cheating?
That’s between you and whatever game you’re playing. It’s built as a helper, not an autoplay button: it narrows the list of words consistent with your clues and suggests a reasonable next guess, but it doesn’t play the game for you or reveal the answer directly. Plenty of people use it after they’re stuck, to check their own reasoning, or just to see how the logic works — all of those are exactly what it’s for.
Why does the suggested word sometimes look like an odd choice?
The suggestion is a frequency heuristic: among the words still in play, it scores each candidate by how common its letters are (by position and overall) and picks the highest-scoring one. That favours common, varied letters, which is usually a good way to eliminate candidates fast — but it is not a full information-theory or minimax solver that simulates every possible outcome. It will occasionally suggest a word that isn’t the mathematically optimal next guess, just a solidly good one.
Does the word list match exactly what a specific game would accept as an answer?
No, and this matters. This tool filters a general English dictionary of roughly 8,500 five-letter words — it has no idea what today’s answer is in any particular game, because it isn’t connected to one. A specific word game usually draws its solutions from its own smaller, curated list, so this tool can show a valid English word as a candidate that a given game would never actually use as an answer. Treat the results as “words consistent with your clues,” not as a guarantee of the game’s exact answer set.
What's a good first guess?
Before you have any clues, the best opening guess is simply a word with five different, common letters — no repeats, so every letter you type teaches you something. Letters like E, A, R, S, T, O, N, I and L appear in a large share of five-letter English words, so a word built mostly from those tends to split the remaining list well. There’s no single “correct” starter, and once you have any letters confirmed you’re usually better off letting the suggestion here work from the narrowed list instead.
Can I add more than one guess row?
Yes. Use “Add guess row” for every guess you’ve made so far — the candidate list and suggestion update from all of your rows combined, not just the most recent one. Each row’s colours are read independently and merged, so a green in row one and a grey for the same letter in row three both count.
Does clicking a tile change its letter?
No. Typing sets the letter; clicking the tile (or pressing space or enter while it’s focused) cycles its colour through grey, yellow and green. The two actions are separate on purpose, so you can fix a colour without having to retype the letter.
The list says zero candidates — what went wrong?
Almost always a mis-clicked tile: a letter marked grey in one row that’s actually green or yellow in another, or a colour cycled one click too far. Since grey caps a letter’s count and green/yellow raises it, one contradictory tile is enough to leave no word satisfying every row at once. Re-check each row against what the game actually showed you, starting with any letter that appears more than once in a guess.
Is anything I type sent anywhere?
No server call is involved in the filtering: the word list loads once with the page, and every guess you enter is compared against it in your own browser. What you type into this tool never leaves your device to do the actual filtering — there’s no account, nothing to install, and results appear instantly with no round-trip.
Use cases & limitations
Where this is genuinely useful
- Sanity-checking a guess. Before you commit to a guess, check it isn’t already ruled out by an earlier clue you may have misread.
- Learning the duplicate-letter rule. It’s one of the most commonly misunderstood parts of these games; typing a real example in and watching the count-based logic work is a fast way to internalise it.
- Breaking a genuine stall. When you’re down to a handful of guesses and can’t think of a word that fits every clue at once, the candidate list does that search for you.
- Practising strategy. Comparing the suggested guess against your own instinct is a reasonable way to think about letter frequency and information value.
What it can't do
- It doesn’t know today’s answer. There is no connection to any specific game or its daily puzzle — only the clue logic and a general dictionary.
- The suggestion isn’t a guaranteed-optimal move. It’s a fast frequency heuristic, not a full solved-game search.
- The dictionary is general-purpose, not a specific game’s answer list. Expect the occasional valid English word that a given puzzle would never actually use.
- It can’t catch a mis-entered clue. If a tile is coloured wrong, the filtered list will be wrong in exactly the way that colour implies — the logic trusts what you tell it.