how can I tell I am playing a bot?

Instant replies at any hour, a stock avatar, canned chat, move times with no variance, and a rating that drifts just enough to keep you matched. If you send an off-script message and get either silence or a generic reply, that is a strong tell. The Washington Post piece on Scrabble GO bots documented most of these signals.

are all AI opponents cheaters?

No. A clearly-labelled AI opponent is a practice partner, not a cheater. The problem is when a word game puts an AI into the human matchmaking queue and does not tell you. WordSalvo's AI lives on a separate screen and is marked AI everywhere — pick it on purpose or do not see it.

what do I do if I suspect my opponent is cheating with a solver?

In WordSalvo, report the game — the report flows into the moderation queue and the game's fair-play verdict is cross-referenced. You can also block or mute that account so they never match with you again. In other apps, check the in-app reporting flow; some do not have one.

does WordSalvo ever use bots to fill the queue?

No. Matchmaking is a live queue of real signed-in accounts inside a ±200 rating-point window. If there is no match within the window, you wait. We would rather you wait 30 seconds than be quietly handed a script.

has WordSalvo ever banned anyone for cheating?

The fair-play system is built around suspicious-pattern checks for PvP games. Enforcement details should follow the shipped moderation flow; the important public promise is that suspicious play is treated as a trust problem, not ignored.

is Wordfeud actually safer than Scrabble GO?

On the specific question of bot fill, there is no published evidence either way for Wordfeud, while Scrabble GO has been the subject of multiple independent reports of suspected bots. That makes Wordfeud the safer bet for "real opponents" today, with the caveat that Wordfeud has not publicly detailed cheat detection.

Word games without bots: a fair-play comparison

Who you are actually playing, and who is actually watching for cheaters.

Last updated April 17, 2026 · By Kurt Bijl

In short: If you want a word game that queues you against real humans and checks PvP games for suspicious solver-like patterns, two options hold up: WordSalvo (no bot fill, server-side suspicious-pattern checks) and Wordfeud (no public reports of bot fill, though no published cheat-detection policy either). Scrabble GO and Words With Friends 2 have both been documented by independent press as mixing bots into their matchmaking.

Two empty tile racks facing each other across a 15×15 word-game board, each with the same seven blank tile slots and no visual indication of who the player is.

What counts as a bot

A labelled AI opponent is not a bot in the dishonest sense — you picked it. A bot here means a computer opponent presented as a human: a face, a name, a friendly greeting, and move patterns tuned to keep you hooked. When a word game fills its queue with these, you think you just beat a retiree in Ohio. You beat a script with a stock portrait.

Two separate behaviours get labelled "cheating" in this genre, and it helps to keep them apart. One is lobby-padding — the app inventing opponents so you always get an instant match. The other is engine assistance — a real human opponent pasting your rack into a solver like Scrabulizer or Wordfeud Cheat. A fair word game has to handle both.

How to tell you were playing a bot

Bots in these apps share tells. They respond within seconds at 3am. They never misclick. They never chat beyond canned phrases. Their ratings drift up and down in a narrow band that keeps you within reach. In the 2021 Washington Post piece "My favorite online Scrabble opponent had a secret: He was a robot", a player realised after months of matches that his "friend" was a Scrabble GO bot.

Community detection gets surprisingly good. Words With Friends 2 players on Reddit maintain a running hotlist of suspected bot profiles — Vision Times catalogued names like Tara McCluskey, Carlita Lopez, and Alexa Dimitrov in 2022. A Zynga spokesperson told the same outlet that bots exist "to make sure there are always fresh players available and prevent queue times." If the company has to explain it to a journalist, the in-app UI did not.

Scatter plot showing a baseline cloud of move-quality points with a single outlier dot sitting far above the rest — the kind of pattern a fair-play engine flags for review.

How WordSalvo handles it

WordSalvo's matchmaking queue is a Firestore-backed list of real signed-in accounts matched inside a ±200 rating-point window. There is no bot fill. AI opponents exist at easy, medium, hard, and expert — but you pick one from a separate screen, and every card, chat surface, and result screen marks them "AI". We do not put AI opponents into the human queue.

On top of that, completed online PvP games can run through a server-side fair-play pipeline. The `fairPlayAnalysis` Cloud Function compares suspicious play signals against the player's own baseline and writes a verdict clients cannot read. The point is suspicious-pattern review: solver-like streaks, unusual timing consistency, and quality spikes that do not match the player's normal games. Read more on our fair-play approach.

How Wordfeud handles it

Wordfeud (by Bertheussen IT) has been running since 2010 and we could not find independent reporting of bot-padding in its matchmaking. That is not the same as a public audit — it is absence of evidence — but it is a stronger position than Scrabble GO or WWF2 can currently point to.

Wordfeud does not, as far as we can tell, publish a detailed cheat-detection policy. The Quora thread on detecting cheating in Wordfeud is almost entirely about spotting humans using external solvers — the same solvers that openly advertise against Wordfeud. If Wordfeud runs per-move statistical analysis, they have not said so publicly.

How Scrabble GO and Words With Friends 2 have been criticised

Scrabble GO (Scopely) has been the subject of sustained independent reporting on suspected bot fill. The 2021 Washington Post profile is the most-cited example; analyst Josh Bernoff has catalogued the broader Scrabble GO experience; and Social Catfish documents user complaints of bots presented as humans. Scopely's official help center lists opponent types but does not disclose bot fill in the general pool.

Words With Friends 2 (Zynga) has been in similar waters. Vision Times' 2022 investigation "When Words With Friends Devolved Into Words With Bots" includes Zynga's own acknowledgement to press that bots have been in the game since 2019. Zynga's unrelated 2019 data breach — covered by Sophos — is a separate issue, but it does colour the "trust us" posture when it comes to undisclosed bots.

At a glance

Four word games, four answers to the same three questions. The comparison is conservative — where a claim depends on public reporting rather than a formal audit, we say so in the row.

Bot fill, AI labelling, and fair-play checks across four word games (as of April 2026).

Game	Bots in online queue?	AI opponents labelled?	Published PvP checks	Source
WordSalvo	No. Real matched accounts only, ±200 rating window.	Yes. AI is on a separate screen and every surface marks it "AI".	Yes. `fairPlayAnalysis` suspicious-pattern checks, server-side, verdicts clients cannot read.	Fair-play approach; FAQ
Wordfeud	No public reports of bot fill.	Solo mode exists and is clearly marked.	No public cheat-detection policy we could find.	Quora discussion
Scrabble GO	Suspected by independent press; not disclosed in-app.	Solo mode labelled; general queue bot disclosure not public.	Not publicly documented.	Washington Post, 2021; Scopely help center
Words With Friends 2	Zynga told press bots have been in the game since 2019.	Solo mode labelled; general queue bot presence not disclosed in-app.	Not publicly documented.	Vision Times, 2022

What you can do right now

If you are currently in a word game and suspect the opponent is a bot, five things help. One: message them something off-script — a bot will either ignore or respond with a canned phrase. Two: check their profile for always-online status and a stock portrait. Three: look at their move times — a human has variance; a bot often does not. Four: check their rating graph if the app exposes one. Five: when you are done, pick a game where matchmaking is against signed-in human accounts and fair-play checks are documented. That is a short list, and WordSalvo is on it.

Frequently asked questions

how can I tell I am playing a bot?: Instant replies at any hour, a stock avatar, canned chat, move times with no variance, and a rating that drifts just enough to keep you matched. If you send an off-script message and get either silence or a generic reply, that is a strong tell. The Washington Post piece on Scrabble GO bots documented most of these signals.
are all AI opponents cheaters?: No. A clearly-labelled AI opponent is a practice partner, not a cheater. The problem is when a word game puts an AI into the human matchmaking queue and does not tell you. WordSalvo's AI lives on a separate screen and is marked AI everywhere — pick it on purpose or do not see it.
what do I do if I suspect my opponent is cheating with a solver?: In WordSalvo, report the game — the report flows into the moderation queue and the game's fair-play verdict is cross-referenced. You can also block or mute that account so they never match with you again. In other apps, check the in-app reporting flow; some do not have one.
does WordSalvo ever use bots to fill the queue?: No. Matchmaking is a live queue of real signed-in accounts inside a ±200 rating-point window. If there is no match within the window, you wait. We would rather you wait 30 seconds than be quietly handed a script.
has WordSalvo ever banned anyone for cheating?: The fair-play system is built around suspicious-pattern checks for PvP games. Enforcement details should follow the shipped moderation flow; the important public promise is that suspicious play is treated as a trust problem, not ignored.
is Wordfeud actually safer than Scrabble GO?: On the specific question of bot fill, there is no published evidence either way for Wordfeud, while Scrabble GO has been the subject of multiple independent reports of suspected bots. That makes Wordfeud the safer bet for "real opponents" today, with the caveat that Wordfeud has not publicly detailed cheat detection.