Word Chums Rack and Board Solver

Word Chums is a turn-based word game published by PeopleFun. Players take turns placing letter tiles on a board, and the highest score when the game ends wins. The board is 15 squares by 15, a size reported by The Word Finder's Word Chums board guide and by Last Letter First's review of the game. A turn starts with seven tiles, which is the number both the review and PeopleFun's scoring help page use.

How a Turn in Word Chums Works

The placement rules matter to a solver because they explain why a list of words is not a list of plays. The first word of a game must cross the center star. After that, every word you form must touch a letter that was already on the board. All newly placed letters have to sit in one row or one column. If a new tile touches letters in both directions, both the horizontal and the vertical word have to be valid. PeopleFun's basic rules page states each of these requirements directly.

Each turn also comes with tools that change what a good move looks like. A hint marks one square that can be used for a good word, and you may use one hint per turn. A bomb replaces your rack with seven new tiles, also once per turn. Pass skips the turn, and three consecutive passes end the game. Swap and Pass lets you exchange selected tiles and then hands the turn to your opponent. Resign ends the game and counts as a loss. The same help page documents each option.

A blank stands in for a letter you do not have; The Word Finder's Word Chums guide says blanks can stand in for any letter and score zero. The game ends when one player uses all of their letters and there are none left to draw, or when players pass three times in a row. In a single-player game, the player who goes out receives the point value of the letters left in the opponent's rack, and those points are subtracted from the opponent's score. The multiplayer and team versions settle the same way with more racks involved.

That rule matters at every stage. The word you play is only part of the turn. What you leave on your rack, and what your opponent can do with the space you open, decide a large share of close games.

The Scoring Rules That Affect Word Choice

Word Chums scores a turn in layers. PeopleFun's scoring help says each letter carries a point value printed in the lower left corner of its tile, and a blank is worth zero. Letter multipliers add value before the word is scored: the help page lists L2, L3 and L4 squares. A letter multiplier applies only the first time a letter is placed on that square. After the letters are valued, the word's total is multiplied by any word multiplier it covers. It lists W2 and W3 squares, and notes that multiple word multipliers compound: a W2 and a W3 produce a 6x multiplier.

Cross-words are scored separately. If a newly placed letter forms a word vertically as well as horizontally, that letter is counted in both words, and both words use the squares they cover. That is why a rack solver cannot rank plays without the board.

Then there is the long-play bonus, where Word Chums departs from the game it is most often compared to. PeopleFun's table reads: five letters plus 10 points, six letters plus 20 points, seven letters plus 40 points. Hanging Hyena's Word Chums page independently lists the same three figures. Placing seven letters is still the largest single bonus, but placing five already earns one. Length starts to matter at five, not only at seven.

The game also rates each word while you build it: the rules help page says two stars means at least 25 percent of the best score available that turn, three means 50 percent, and the point badge turns green at 50. That feedback compares your word with the whole board, not just with the other words on your rack.

How the Board Solver Works

The board tool uses the same dictionary as the rack tool but applies the placement rules instead of ignoring them. You describe the position once, on a grid of 225 squares, and the tool then answers a different question: not which words your letters can spell, but which of those words can actually be laid on this board without breaking a rule or inventing a word at right angles.

The rules it enforces come from PeopleFun's own help pages. The first word of a game must cross the center star. Every word after that must have at least one letter adjacent to a letter played in a prior turn. All newly placed letters must be adjacent to another letter and sit on the same vertical or horizontal row. And every tile placed must form valid words vertically and horizontally if it is adjacent to more than one letter already on the board. The basic rules page states each of those requirements in one short paragraph.

The last requirement is the one a rack solver cannot handle and the one that produces most of the surprise in a real game. A five-letter word dropped into a gap may be legal as a word and still be illegal as a play, because the tile in the middle of it also forms a two-letter or three-letter word running down the board. The board tool reads that perpendicular line for every square it fills and keeps the play only when the line is a word in its dictionary. If the perpendicular line is a single tile with empty squares on both sides, no cross-word exists and there is nothing to check.

A turn is modeled as one unbroken run of letters in one row or one column, which is the shape those rules describe. Existing letters inside the run count toward the main word and are not charged to your tray. The run is also extended to the end of the letters already on the board, so the tool will not offer a five-letter word when the squares on either side hold letters that would really make it seven. Every empty square between the first and last new tile has to be filled by the play, because a gap would leave a placed tile outside the word it belongs to.

Tray accounting is strict, duplicates included. The tool counts the letters of the tiles you place, not the whole word, so a hook that adds one letter to a six-letter word already on the board costs one tile. A question mark in the tray covers one missing copy of a letter and is spent in the process. A play that needs a second B when the tray holds one B and no blank is discarded rather than shown with a warning, because a list of plays you cannot make is worse than a shorter list.

The touch rule is applied in both directions. On a board with at least one letter, every play offered includes a square that is already filled; a word placed in an empty corner with no contact is not a legal turn, so it is not offered. On an empty board the tool switches to opening mode and only offers placements that cover the center star, which is the one case where contact with an existing letter is impossible.

The search is exhaustive within those constraints. For each empty square and each direction, it considers every start position and length that could work, then tests candidate words from a per-length index. The index is built once when the dictionary loads, so pressing the button does not re-read the 2.7-megabyte word file. A full board is nearly instant because every position is blocked. A crowded board with many crossing points and a large tray is the slow case, and on this dictionary it is measured in tenths of a second, which is why the search runs on a button press rather than on every keystroke.

Reading a Play List

Each line gives four things: the word, its direction, its starting square, and the tray letters it uses. The starting square is the first letter of the word, numbered from the top left corner of the board, so row 1 is the top edge and column 1 is the left edge. A play shown as STAINER, across, row 8 column 4, occupies row 8 from column 4 to column 10. If three of those squares already held letters, the tray line lists only the four tiles you would place.

Direction matters more than it looks. The same word can be worth very different amounts across and down, because it covers different squares, and the cross-words it creates are different words entirely. The tool prints both directions as separate plays whenever both are possible, and it does not collapse them into one line.

Plays are ordered by the length of the main word, longest first, and alphabetically inside each length. That order is a deliberate admission. A true ranking would need a score for every play, and a score needs the tile values plus the premium-square layout, neither of which this page can verify. Length is at least a fact: it comes from the dictionary and from your tray, and it decides the long-play bonus, which PeopleFun publishes. Within a length, alphabetical order is neutral. It does not pretend that BLEACHER is better than TREACHER, and it lets you scan for a prefix or a suffix you want to keep.

The list shows the first two hundred plays and prints the true total for the position. The show-all button renders the rest. Two hundred is a reading limit, not a search limit: the count in the summary is everything the tool found, so a large number is information about the position rather than a bug. An empty board with AEINRST returns 2,916 plays, 154 of them seven letters. Almost all of the rest are short words in corners you would not choose, which is why the length order and the cap both exist.

A worked example shows the difference between a word list and a play list. Put CAT on row 8 in columns 1 to 3 and give the tool a single S. The rack box cannot help with one tile, because it is built to spell words from a whole rack. The board tool returns exactly three plays: CATS across from row 8 column 1, AS down from row 8 column 2, and ST down from row 7 column 3. Two of the three are one-tile hooks onto a letter that was already there, which is the kind of play a word list will never surface.

Nothing is deduplicated across squares. A word that fits in four places appears four times, because those are four different plays with different cross-words and different consequences. If you only want the word itself, the rack tool on this page gives you the word list without the placements.

Choosing among the plays still needs judgment, and the judgment is about things the tool can see and things it cannot. It can see length, so start with the longest group and the bonus table. It can see the letters you would spend, so prefer a play that leaves a balanced rack over one that leaves four vowels. It cannot see the score, so look at the board yourself: a short word that covers a word multiplier you can identify on screen may beat a long word in the open, and a play that opens a lane to a premium square for your opponent may be worse than a quiet one. The tool's job is to make sure every candidate it shows is a real play, not to decide which one wins the turn.

Entering a Board Accurately

Paste mode is faster for a board you are copying from a screenshot or from another device. Type or paste up to fifteen lines of fifteen characters, using a letter for a filled square and a dot, dash, asterisk, question mark or space for an empty one. Rows shorter than fifteen are padded with empty squares, and fewer than fifteen lines is accepted, which makes a half-transcribed board usable while you finish it. A row longer than fifteen characters is refused with the row number, because trimming it silently would shift every letter after the cut by one square and produce confident nonsense.

The grid is the correction tool. It is a keyboard grid of 225 single-character boxes with the row and column numbers printed along the edges, and it edits the same board that paste mode loads. Typing a letter fills the square and moves the cursor one square to the right, wrapping to the next row at the edge. Backspace clears the current square and steps back; pressing it on an empty square steps back and clears that one. The arrow keys move without changing anything, so you can navigate to a specific square and type over it. Enter runs the search from anywhere in the grid, and the Clear board button empties every square.

Three transcription errors account for most bad results. The first is a row shifted by one square, which happens when a leading space is dropped or added; the grid makes it visible because the letters no longer line up with the numbers. The second is a missing letter in the middle of a word, which turns a real word into a gap the solver will happily fill with a tile you do not have. The third is a played blank. A blank that is already on the board should be entered as the letter it now stands for, because that is what it counts as for every word it touches; entering a question mark there tells the tool the square is empty.

The center star is worth checking before you trust an opening search. On an empty board the tool only offers words that cover row 8, column 8, and it labels the square with a star so you can confirm the grid is oriented the way the game is. If your board is rotated, every placement will be wrong in a way that looks plausible, which is the worst kind of wrong. Row 1 at the top and column 1 on the left is the convention this page uses everywhere, including in the play list.

Nothing is uploaded. The board exists as a grid of characters in this browser tab, the dictionary is a file served from this site, and the search runs on your device. There is no account, no saved board, and no image upload, so a screenshot has to be read by you and typed in. That is a limitation, and it is also why the tool works with no network connection once the page has loaded.

What This Rack Solver Does

The tool treats your input as a hard inventory. It counts each letter, respects duplicates, and only returns a word if every copy of every letter in that word is either in your rack or covered by a blank. That is the difference between a word finder and a rack solver: a page that returns words merely containing your letters will offer plays you cannot make.

The page loads the site's general dictionary file, which contains 267,751 space-separated entries. After dropping entries shorter than two letters or longer than twelve, 239,021 words remain. That count is measured from the file itself, not estimated. The file is split and grouped by length once when the page loads. Each search then scans only the length groups that could fit, which is why typing a new letter does not trigger another parse of a 2.7-megabyte file.

Results are grouped by length, longest first, and sorted alphabetically inside each group. Every group shows its own count, and every group has a copy button. There is also a copy-all button for the full result set. Nothing is uploaded: the dictionary and the search code are files served from this site, and the letters you type stay in the browser.

The rack box is deliberately narrower than the board tool. It returns candidate words, not placements; which word fits a particular opening is still a board question, and the board tool above is where that question gets answered. Keeping the narrow tool is still worth doing, because a word list with no board attached is the right answer when you are studying a rack or checking whether a swap makes sense. The Word Finder's board tool takes the placement-first approach and asks you to recreate the board around a play.

How to Use the Rack Solver

1. Type the rack exactly as it appears. Include repeated letters. AABDELS and ABDELS are different racks: the first can build ABASED, the second cannot, because it has only one A.
2. Use a question mark for each blank. One ? per blank. The solver tries every letter and keeps substitutions that produce a word.
3. Add a filter only when it saves reading. Starts with, contains in order, and ends with accept letter sequences. Exact length restricts the output to one group, which is useful when the board has room for exactly five letters.
4. Use the hook field for one board letter. Type the single letter that is already on the board and choose whether it must be first, last, inside, or anywhere in the word. The solver adds one copy, because the board supplies it, and does not remove a rack tile.
5. Read the longest group first. In Word Chums the bonus starts at five letters and reaches its maximum at seven, so check seven first, then six and five.
6. Copy what you need. A group copy gives you that group's words. Copy all adds headings. The buttons use the clipboard when available and fall back to a temporary text field.

A worked example shows the output. Enter AEINRST and the solver returns 343 words. Eleven of them are seven letters: ANESTRI, ANTSIER, NASTIER, RATINES, RESIANT, RETAINS, RETINAS, RETSINA, STAINER, STARNIE and STEARIN. Several are obscure, which is the point of the list: it does not care whether you recognize the word, only whether the letters can spell it.

Now add a blank. BA?TERS returns 2,102 words, 48 of them seven letters, including BANTERS, BEATERS, BOATERS, BLASTER, BRAVEST and BREATHS. Without the blank, the same six letters cannot reach seven at all. The blank is doing real work.

Hooks change the shape of the list. With AEINRST and a board L, the solver returns 312 candidates. Restricted to words that start with L, the count is 99, including LATRINE, LINTERS, LARNIEST and LATRINES. Restricted to words that end with L, the count is 64, including ENTAIL, STERNAL and REINSTAL. Those totals come from this dictionary; another word list will differ.

Reading the Results by Length

The groups are ordered longest first because the game rewards length. A turn that places seven letters earns the 40-point bonus; six earns 20, and five earns 10. A seven-letter word deserves the first look. If the seven-letter group is empty, the six-letter group is not a consolation prize; it is the next tier of the same bonus structure.

Inside a group, the words are alphabetical. Without verified tile values, any other order would be a guess dressed up as a ranking. Alphabetical order lets you scan for a prefix, suffix, or a letter you want to keep. If you need a particular shape, the filters are faster than reading: set the exact length, add the prefix or suffix, and the group shrinks to something you can judge.

The count is useful. A group of three words is worth reading in full. A very large group is not, and the exact-length filter or a prefix is the right response. The count also tells you whether a rack is genuinely productive. A seven-letter rack that produces no five-letter word is usually a weak rack even if it produces a long list of three-letter words, because the bonus rewards longer plays.

Length is still not the same as score. A four-letter word placed across a triple-word square can beat a seven-letter word played into the open, and a play that forms two or three cross-words can beat both. The solver gives you the words; the board decides which one is worth playing. Treat the length groups as a shortlist, not a ranking.

When the Rack Is Bad

A bad rack is one with no play that improves your position: no five-letter word, no useful hook, and no way to shed the tiles that are clogging your next draw. Almost every rack has words, so the solver is good at identifying the first part of that description and honest about not knowing the rest.

Four patterns cover most bad racks. The first is a rack with no five-letter word, so the bonus is not available from the rack alone and the best play is probably a short hook. The second is a duplicate-heavy rack: three Es, two Os, or a pair of Us. Duplicates narrow the list and usually point toward a short word rather than a long one. The third is an all-vowel or all-consonant rack, the situation the bomb is designed for. The fourth is a rack built around a premium letter with no home, usually a Q with no U and no short rescue.

What you do next depends on the board and the score, not just the rack. A short word that keeps a balanced rack can be better than a longer word that leaves you with five vowels. Swapping tiles costs a turn but can turn a dead rack into a live one. A bomb draws seven new tiles, one per turn. A hint marks a square that can support a good word, also one per turn. None of those decisions is visible to a rack solver.

One useful habit is to run the rack before you swap. A hopeless-looking rack often contains one five-letter word the eye missed. A long list of two- and three-letter words and nothing longer is evidence for swapping. If the solver returns nothing, the rack cannot form any word in this dictionary, a strong signal to exchange.

Blanks deserve their own note. A blank scores zero, so using one in a short word buys position rather than points. That can be correct when the play blocks a premium square, but it is usually a poor trade when the blank could help you place five or more letters in one turn later. The solver shows every word the blank makes possible and leaves that judgment to you.

Hook Letters and Board Connections

A hook is a letter or short word attached to a play that is already on the board. In Word Chums, every word after the first must connect to an existing letter, so hooks are not an advanced trick; they are the normal way the board grows. The solver's hook field exists to narrow a long list to the words that contain the letter you intend to build on.

The mechanics are simple: type one letter and choose whether it must be first, last, inside, or anywhere. The solver adds one copy of that letter to your available tiles and then applies the normal buildability test. If you hold two L tiles and the board has an L, three Ls are possible; the board copy is counted separately. If you leave the hook field empty, the position selector has no effect.

Anywhere is the honest default, because a connecting letter can sit at the start, end, or middle. First and last are useful when extending a word in a line; inside is for crossing at right angles. None of these settings proves that the placement is legal. They only prove that the word has the right shape.

The rack tool's hook filter does not know which squares are open, whether the new letters are adjacent to the existing word, or whether the perpendicular words are valid. It also does not know which premium squares are still reachable. A word that passes the hook filter can still be impossible to place, and a word that fails the filter can sometimes be played by attaching a different letter. The filter reduces reading; it does not replace the board. The board tool above is the part of the page that does check adjacency and cross-words, and it does that from a position you type in rather than from the rack.

Three questions settle most placement choices. Can the word physically reach a premium square from an open lane? Does the placement create a cross-word that adds points? What does the play leave for your opponent? The hook filter helps with the first only indirectly, by giving you words that contain the letter you want to build on. The other two are board questions, and the board tool at the top of the page answers the second one for you.

Why There Is No Score Column

A score column would be convenient, but it would also stop you thinking about tile values. PeopleFun's scoring help says every letter has a point value printed in the lower left corner of the tile, and that blanks are worth zero. It lists L2, L3 and L4 letter squares and W2 and W3 word squares. It does not publish a complete letter-value table on the pages we read.

Third-party tables exist, and they do not agree. One published table gives different values for C, J, K and V than another, and each presents its numbers as the game's. We could not check either against the game client, and the help pages we read do not publish a table. Tile values may also differ between versions of the game, so this page does not print them. Printing one table would turn a coin flip into an authoritative-looking column.

Tile values are only half of the problem, and the other half gets less attention. A score also needs to know which square a letter landed on, so it needs the premium-square layout: where the L2, L3, L4, W2 and W3 squares sit on the 15 by 15 grid. PeopleFun's scoring help names all five multiplier types and shows a swatch for each, but the page we read does not print their positions. The only complete map we found sits in the page data of one competitor's board editor, and that tool does not commit to a single board: the editor at wordchumscheat.com offers a Standard Game board and a smaller Quick Game board, each with its own arrangement of multiplier squares. We cannot tell which map matches the version and mode you are playing, and a wrong map would corrupt every score on the page at once.

So the honest position is narrower than it looks. We can say that a blank is worth zero and that five, six and seven letters earn 10, 20 and 40 bonus points, because PeopleFun publishes both facts. We cannot say what a word is worth, because that needs a value for every letter and a map for every square, and neither is published in a form we can verify. A board tool that prints a score is making two unverified assumptions per play and presenting them as one number.

A wrong score is worse than no score. It does not merely fail to help; it changes which word a player chooses, and the player has no way to see the error. The same logic applies to a score column built from another game's values. Word Chums has a four-times letter square, so the board is not a straight copy either.

What the page does instead is print facts that can be checked. Length comes from the dictionary and cannot be wrong. The long-play bonuses come from PeopleFun's own table and are confirmed by a second source, so they are stated with citations. Blanks are worth zero, which is also from the official page. Everything else about scoring is left to the game, where the tile values are visible on the tiles and the multipliers are visible on the board.

If PeopleFun publishes a complete table, or two independent sources agree on one, this page can add scores later. Until then, the output is a list of words and lengths, not a number that looks precise and is not.

Common Mistakes When Using a Rack Solver

• Putting board letters in the rack. The rack box is an inventory of tiles you hold. If you add a letter that is already on the board, the solver will offer words you cannot build. Use the hook field for one board letter, and leave the rest of the board out of the rack.
• Dropping a duplicate. If you hold two Es, type two Es. The solver respects counts, so a rack entered with one E will hide every word that needs two. The ABASED example above is the quickest way to see the difference.
• Typing a blank as a letter. A blank is not the letter you hope it becomes. Enter ? and let the solver test the substitutions. Typing an E instead narrows the search to words that use an E and hides the ones that need an A, an R or a consonant.
• Treating length as score. The bonus table rewards five, six and seven letters, but word and letter multipliers can overturn the order. A short word on a premium square can beat a long word in the open, and a play that creates two cross-words can beat both.
• Assuming every result is legal. This page uses a broad English dictionary, not the game's internal list. Word Chums may reject an obscure entry, and it may accept a word this list lacks. Check doubtful words in the game before you commit them.
• Confusing a prefix filter with a board position. Starts with filters the spelling of the word. It does not mean the word will start at the left edge of the board or attach at the beginning of an existing word. In the rack box, only the hook field deals with a board letter, and even that does not prove placement. Use the board tool above when placement is the question.
• Using the solver during a live game against another person. A solver is a study tool. Consulting one mid-game is generally considered cheating, and it defeats the practice that makes the tool useful in the first place. Use it after the game, or in solo practice.

Dictionary and Word Legality

The word list behind this page is a broad English dictionary, not PeopleFun's internal list. That distinction matters more than any other limitation on this page. The Word Finder's history of Word Chums notes that the game uses its own dictionary, so a word's status can differ in both directions: the game may reject something shown here, and it may accept something this list does not contain.

The risk is highest at the edges of the language. Short entries, dialect forms, unusual inflections, and borrowed words are exactly where lists diverge. A solver can only tell you what its own list contains, which is what this page does.

If the game rejects a word, the solver is wrong for that board, not the game. The final authority is the game's own acceptance. That is not a defect in the tool; it is the boundary between a general dictionary and a game-specific one. Building a fake Word Chums list from a general list plus guesswork would be worse than admitting the boundary, because it would hide the uncertainty instead of exposing it.

The list does have practical advantages. It surfaces obscure words, covers two through twelve letters, and is local, so it works offline once the page has loaded. If you want a general word finder for a different game, the same list powers the site's word finder and anagram solver. They use the same vocabulary but answer different questions.

Practice, Fair Play, and When to Close the Tab

A rack solver is most valuable between games. Deal yourself seven tiles, give yourself sixty seconds to find the best play, then run the rack and compare. The gap between your answer and the solver's list is the lesson. Ten rounds will show the patterns that produce five-letter words: common suffixes, vowel-consonant balance, and hook letters. None of that requires a live opponent or gives you an advantage in a current game.

Using the tool during a live game against another person is a different matter. It is generally considered cheating, and it undermines the practice that makes the tool worth using. The same applies to a ranked match, a tournament, or any game where the other player believes you are choosing words yourself. If you want to check whether a word was legal after the game, that is study. If you want to find a word while the clock is running, that is not.

The solver is also useful for reviewing a finished game. Take the rack you held on a turn that went badly, enter it, and look at the groups you did not consider. You will often find a five-letter word that was available and a shorter hook that would have kept the rack balanced. The review is more instructive than an in-game answer because you can see the board and the score.

How This Page Compares With Other Word Chums Tools

Two established tools come up often for this game, and both are worth knowing about. wordchumscheat.com presents a rack field with contains, begins-with and ends-with filters, accepts question marks for blanks, and lets you choose 10, 25, 100 or all results. It also runs a separate board editor, which is a reasonable choice if the board is your question.

hanginghyena.com/wordchumscheat takes a rack plus a start sequence, an end sequence and an includes field that uses one letter from an existing word. It ranks results by points and says its scores use current Word Chums tile values. It also explains the bonuses for five, six and seven letters played. When we checked in October 2026, its footer read copyright 2012 to 2022, so the page has not been refreshed in some years. It has no board input at all.

Board editors, side by side

wordchumscheat.com's board editor is the tool to beat on this page, so it is worth describing exactly what it does, from a visit in October 2026. It presents a full board you fill by tapping squares, with a keyboard underneath. Its instructions say you set the typing direction by swiping a square vertically, or by holding control and an arrow key on a desktop, and that a question mark or a period enters a blank; the page's own prompt asks you to identify a tile it could not read, which points to an image-based import path as well as manual entry. It offers a Standard Game board and a Quick Game board, keeps named boards you can save, load and delete, and its page text says the editor is there "to see the highest-scoring moves you can make", so it ranks by score.

This page's board input makes a different trade. There is no tile recognition and no saved board, so it cannot import a screenshot; the trade is that there is nothing to load, no dialog to dismiss and no account, and the whole thing works offline once the page is cached. Entry is either a pasted block of fifteen lines or a keyboard grid where typing advances, Backspace steps back and the arrow keys navigate, which is faster than tapping 225 squares on a desktop. The output is the difference that matters most: this page prints the direction and the starting square for every play, checks the cross-words before showing anything, and prints no score. A tool that prints a number has to be right about the tile values and the square layout. A tool that prints a length and a coordinate only has to be right about the dictionary, which is a claim this page can support.

Where the other tool is stronger is convenience. If you already have a screenshot and you want a scored answer, its board editor does that in one pass, and this page will not. If you want to know whether a specific placement is legal, and roughly where the long words go, this page answers without asking you to trust a number it cannot verify. The two tools are solving the same problem with different definitions of "best".

This page also differs from both competitors in its rack mode. It groups results by length and prints a count per group, which makes a large result set easier to navigate. It offers an exact-length filter and a hook-position filter, neither of which appeared in those two tools. It accepts up to twelve characters for testing. And it deliberately omits scores, for the reasons explained above.

One limit is now shared only by the competitors: a rack-only tool cannot see the board at all. They can rank words or score words, but the placement decision stays with the player. This page closes that gap with the board tool above, and inherits the next limit in exchange, which is that the board has to be entered correctly. A board editor is only as good as the position you type into it, which is why the grid prints row and column numbers and why the paste parser refuses a row that is too long instead of trimming it.

What This Solver Cannot Do

The rack box cannot see the board. It does not know which squares are occupied, which premium squares are open, or which words are already down. It cannot check adjacency, and it cannot validate the cross-words a play would create. The hook field is a spelling filter, not a placement engine. That is what the board tool above is for, and the two are separate on purpose: a word list and a placement list answer different questions.

Neither tool can score a play. There is no verified tile-value table and no verified premium-square layout, so the page does not pretend to know a word's score. The published long-play bonuses are stated separately because they can be cited. Everything else about scoring is left to the game, where the tile values and the multipliers are visible on screen.

It cannot read a screenshot. There is no image upload and no optical character recognition, which is one place a competitor's board editor goes further. Typing a board is slower than photographing it and considerably more reliable, and a grid you can see is easier to correct than a recognition result you have to audit.

It cannot guarantee that a word is legal in Word Chums. The dictionary is broad and general, and the game's list is its own, so a play shown here can still be rejected by the game and a play the game accepts can be missing. It cannot simulate a hint or a bomb, track the tiles left in the bag, or tell you what your opponent is holding. It cannot tell you which premium squares are still live, because it does not know the layout. Those are game-state questions.

What it can do is answer two questions completely and quickly. Given these letters, which words can be built, how long are they, and how many are there. And given this position, which of those words can be placed, in which direction, starting where, without breaking a rule or inventing a cross-word. Both answers are checkable against the dictionary the page loaded, which is the only authority it claims.

Sources Checked

• PeopleFun Support, What are the Basic Rules for Word Chums? Used for placement, cross-words, adjacency, the center star, hints, bombs, passes, swap and resign.
• PeopleFun Support, How does scoring and winning work? Used for tile values, blanks, L2/L3/L4 and W2/W3, bonuses, stars and endgame scoring, and for the absence of a published square layout.
• The Word Finder, Word Chums Board Solver, The History of Word Chums and Word Chums Cheat. Used for the 15 by 15 board, the game's own dictionary, the blank rule and one of the conflicting value tables.
• Last Letter First, Word Chums vs Last Letter First. Used for the 15 by 15 grid and the seven-tile rack.
• Adweek, Word Chums mobile game review. Used for the publisher and the seven-letter turn.
• Hanging Hyena, Word Chums Cheat. Used for the bonus figures and the tool comparison.
• Word Chums Cheat, Word Chums Cheat and Solver and its Board Editor with the board editor tutorial. Used for the tool comparison, the typing-direction and blank-entry conventions, the two board sizes and the saved-board feature.
• Word Unscrambler, Word Chums Cheat. Used for the conflicting third-party value table and the tool comparison.