A word search is unusually hostile to an unconventional reader: dense, low-redundancy text with no context to fall back on. Larger letters, higher contrast and a smaller grid fix most of it. Reversed words and confusable letterforms are the parts no setting can fix.
Most apps become accessible by making the text bigger. A word search does not get off that lightly.
Ordinary reading is enormously redundant. If a word is slightly blurred you recover it from the sentence; if a letter is ambiguous you resolve it from the word. A grid strips all of that away. There is no sentence, no meaning, no predictability — just a dense field of unrelated characters that each have to be resolved on their own, at small sizes, with nothing to fall back on.
That makes it one of the least forgiving things you can put on a screen, and it means the usual accessibility advice only gets you part of the way.
What makes a grid hard to read, specifically?
Three things, and they compound.
Density. A 14×14 grid is 196 characters in the space a paragraph would use for perhaps forty. Every letter is smaller than body text would be on the same screen, and they are packed with minimal spacing because the grid has to stay square.
No redundancy. The letters are deliberately meaningless. You cannot guess a smudged character from its neighbours, because its neighbours are random too. Every single cell has to be independently resolved, which is a much heavier perceptual load than reading at the same letter count.
No context to recover from error. Misread one letter in a sentence and you barely notice. Misread one letter in a grid and the word you are hunting for simply is not there.
Those three are why a puzzle that looks trivially simple can be genuinely exhausting, and why the fix has to be structural rather than cosmetic.
What should you change first, and in what order?
Text size, then contrast, then grid size — and the third one is the surprise.
Text size. The obvious one, and the one to be least shy about. In WordHunt letter and interface text scale through four steps. Set it one step past what looks normal; if it turns out to be more than you need, step back. Almost nobody does this, and almost everybody who does finds the puzzle became noticeably easier without becoming less interesting.
Contrast. Underrated, and a separate capacity from sharpness. Contrast sensitivity can be poor in someone whose acuity tests perfectly, which is why mid-grey letters on a pale background feel tiring within minutes even when they are technically legible. A dedicated high-contrast palette raises the separation between letters, grid and highlights, and it is usually the single biggest improvement after size.
Grid size. This is the one people miss. Dropping from a 14×14 grid to a 10×10 one makes every cell substantially larger at the same screen size — a bigger change to the actual letter size than the text-size setting produces, and it costs nothing but a shorter puzzle. If reading the grid is the hard part, play a smaller grid. There is no prize for the bigger board.
The full set of settings is on the features page, and the same ordering applies to the age-related changes we cover in word searches for older players.
Why is a word search particularly awkward with dyslexia?
Because it isolates the exact operation that tends to be effortful, and removes every support for it.
Reading difficulties are varied and individual, but a common thread is that decoding — mapping letters to sounds, discriminating between similar letterforms — takes more deliberate effort than it does for other readers. Fluent reading normally compensates for this with context: meaning, syntax and predictability all help resolve a word that decoding alone would struggle with.
A word search deletes every one of those supports. What remains is pure letterform discrimination, repeated several hundred times per grid, with no meaning to lean on. It is close to a worst-case presentation.
Two specific features make it harder still:
- Confusable letterforms. The b/d/p/q family are rotations and reflections of one shape. In running text, context usually resolves them. In a grid there is no context, and a
bread as adsilently removes a word from existence. - Reversed words. From Medium difficulty upward, words can run right to left. For a reader whose directional processing is already effortful, this is a genuinely unfair addition.
The practical answer is the difficulty ladder rather than a setting. Easy is 8×8, five or six words, running across and down only — no diagonals, no reversals. That is not a beginner's consolation prize; it is the version of the puzzle with the two hardest elements removed, and it is a perfectly good place to stay permanently.
What can a screen reader actually do with a grid?
Nothing at all, unless the cells are individually labelled — and most word games do not label them.
This is the gap worth checking before you commit to an app. Labelling the buttons is easy and nearly everyone does it; labelling several hundred grid cells is more work and many do not bother. The result is a game where a screen-reader user can navigate the menus perfectly and then hit a completely opaque rectangle where the puzzle should be.
In WordHunt every control and grid cell carries a label, so the game is navigable with TalkBack. That does not make a word search an ideal screen-reader experience — it is an inherently visual task, and no amount of labelling changes that — but it does mean the grid is readable rather than a void.
Related, and easier to overlook: no control is smaller than 44 points. On a dense grid that floor does real work. If a drag slips by a millimetre or two, whether through tremor, limited fine motor control or simply walking, a larger target is the difference between selecting the word you found and selecting a diagonal you did not mean. A wrong selection should also cost nothing — no penalty, no lost points, no failure sound. You lift your finger and try again.
Does anything help beyond the settings?
Yes, and it is the same technique that helps everyone, only more so.
Hunt one letter, not the word. Instead of scanning for the shape of PANDA, scan only for P and check each one. This converts a demanding whole-word pattern match into a single-character comparison, which is far less work per cell. Pick the rarest letter in the word rather than the first. It is the highest-return habit in word searching generally, and it helps disproportionately when whole-word recognition is the expensive part — the mechanism is in why you cannot see a word that is right there.
Search for the last letter on reversed words. Rather than fighting the direction, aim the automatic process the other way.
Use the reduced-motion setting if animated transitions make you queasy or make the grid harder to settle on. It honours the system setting and drops the animations.
Stop when it stops being pleasant. A word search is a game. If a grid is costing more effort than it returns, the answer is a smaller grid or an easier difficulty, not persistence.
What can no setting fix?
Two things, and it is worth being straight about them.
The task is irreducibly visual. Labels make a grid navigable with a screen reader, but a word search is a spotting exercise, and no amount of accessibility engineering turns it into something else. For a blind player there are better puzzle formats, and we would rather say so than pretend otherwise.
And the harder difficulties are harder for reasons that are not adjustable. Diagonals, reversals and a 14×14 board are the puzzle, not a presentation choice. The honest position is that the full difficulty range is not equally available to every player, and that the lower rungs are a complete game rather than a waiting room — most people settle around Medium and stay there, which is exactly the right outcome.
Everything above is a setting you change once, in about a minute, and then never think about again. That minute is worth more than any technique in the rest of this blog.
