Wordle Solver

There are three input fields. 'Correct' is for green letters in their known position — for example, if the first letter is C and the third letter is T, type 'C_T__' (underscore for unknown). 'Present' is for yellow letters — letters that exist in the word but not at that position. 'Absent' is for gray letters that have been ruled out.

After each guess, take Wordle's color feedback, transcribe it into the solver, and press Solve. We filter every valid 5-letter word in the dictionary by your constraints. Results are sorted by letter frequency, so the words at the top are statistically the best next guesses (they cover the most still-possible answers).

When you're truly stuck, look at just one or two of the top suggestions; reading the entire list takes away the puzzle. Think of the solver as a safety net rather than an auto-solver — it keeps you from running out of guesses but doesn't play the game for you.

Computer simulations have studied millions of Wordle games. The best starting words by 'average information gained' include SALET, CRATE, SLATE, CARTE, TRACE, and STARE. All of them share two properties: high-frequency consonants (S, T, R, N, L) and at least one common vowel (A or E).

Your second guess should sometimes be a 'sacrifice guess' — a word with completely new letters, even if it has zero green or yellow boxes left to use. This trades a turn for information. NYMPH, GLYPH, and CRYPT are popular sacrifice words because they test rare letters (Y, P, H, M, N, C, R) with no vowels.

Avoid repeating letters too early. If your first guess was STARE and you got an S yellow, your second guess should NOT also start with S. You already know S is in the word; spend the turn testing other letters.

Solving Wordle is essentially an 'information maximisation' problem. Each guess reduces the remaining answer pool by some factor. Mathematically, the best guess at any point is the one that maximally splits the pool of remaining candidates across the possible color outcomes.

Researchers at 3Blue1Brown and elsewhere have computed that the theoretical optimal player can solve Wordle in an average of 3.42 guesses. The average human player solves it in roughly 4.0 guesses. The gap — about half a guess — is mostly about choosing 'high information' starters and being willing to use sacrifice guesses.

Our solver doesn't compute information theory in real time (that would require comparing every candidate against every other candidate, which is computationally expensive), but it sorts results by letter frequency, which approximates the same idea well enough for practical play.

The most common beginner mistake is putting yellow letters into the 'Absent' field. A yellow letter IS in the word; it's just in a different position. It belongs in 'Present', not 'Absent'.

The second common mistake is with duplicate letters. If you guess GEESE and get one E green and the other E gray, the word contains exactly ONE E. Our solver handles this automatically — just enter each colour in its right field and we sort the rest out.

Third is using the solver before your first guess. With no constraints, the solver returns thousands of words — that's not actually useful. Pick a good starter yourself; the solver shines on turns 2–4.