Word Games
The Counterfeit Lexicon
Beside his word-hoard Silas Wren kept a second drawer of forgeries: seven real words, three definitions apiece, and the one true meaning of each sealed before he died.
- Players
- 1
- Length
- 10–15m
- Ages
- All ages
- Version
- 2.0.0
- Updated
Works with ChatGPT · Claude · Gemini. Free, no signup.
Opens ChatGPT or Claude with the play prompt ready; press send if it hasn't started. Same link on your phone as on your desktop. Another app? Copy the play prompt into it. This game needs a chat that can open links: its answers stay in a sealed file on this site. What works where.
No web access, or want it as an installed app?
The game card is every rule of the game, about 10k characters, but not its answers: those stay in a sealed file on this site, so even pasted, The Counterfeit Lexicon needs a chat that can open a link. Pasting the answers in yourself would spoil it. The Skill installs The Counterfeit Lexicon into Claude on any plan: download it, then turn on code execution under Settings → Capabilities and add it at Customize → Skills → + → Upload a skill.
The Counterfeit Lexicon is a free word game you play inside ChatGPT, Claude, or Gemini. A sealed bluffing word game: seven obscure words, three definitions each, the true one hashed before you play. Solo, 10–15m a run. No signup and no app — open it in your assistant, or paste the play prompt anywhere else.
How to start
- Open it in your AI. Tap Open in ChatGPT or Open in Claude, and press send, or paste the play prompt into any assistant that can open links. No paid plan required.
- The game loads. The Curator loads the full card and opens its sealed script from this site, so run it in a chat that can follow a link. It reads you the SHA-256 commitment, then deals Card 1 — a word and three lettered definitions.
- Name the true letter. Read all three definitions and type A, B, or C for the one you believe is genuine. The AI scores it against the sealed answer and reveals the real meaning.
- Verify the seal. At the end, rebuild the SHA-256 hash of the revealed script at aigamecards.com/verify to confirm no answer moved while you played.
Questions
- How do you stop the AI from changing which definition is right?
- The seven true definitions and which lettered slot each sits in were written and locked under a SHA-256 hash published before release, and the Curator reads that hash out to you before the first card. The AI reads the definitions only from that fixed script and cannot move the true answer once you commit to a letter. You can hash the revealed script yourself to confirm nothing changed.
- What makes it different from a trivia quiz?
- The two wrong definitions are not random — they are deliberate forgeries built to sound as plausible as the real one, borrowing real-looking roots and histories. It is a game of spotting the counterfeit, so etymology helps only up to a point.
- Are the words and definitions real?
- Yes. Every headword is a genuine, dictionary-attested English word, and the true definition is its real meaning. Only the two decoys per card are invented. At the end the AI reveals all seven words and what they actually mean.
- Does the chat have to be able to open a link?
- Yes. The Curator fetches the sealed script from aigamecards.com; pasting it in yourself would hand you all seven answers before the first card. Not every app can, and it changes; the compatibility page at aigamecards.com/launch-matrix shows which ones currently do. No account beyond the chat app you already use and no paid plan either way, and the seal is checked afterward at aigamecards.com/verify, in your browser.
- What counts as a good score?
- Four of seven is par. Blind guessing averages a little over two, so anything above that means you are reading the forgeries and not just the words; five or better and you have beaten Wren at his own drawer. There is no partial credit — one letter per card, and it stands — and a card you skip scores nothing. The full seven runs about twelve minutes.
- Has the drawer changed since it was released?
- Once. On 28 September 2026 the Counterfeit Lexicon was re-sealed as a second edition (v2.0.0): seven less-travelled words, and forgeries written in exactly the flat register of a real entry, because the first edition's words were famous enough to recognise and its forgeries read as invented. The first edition's seal (v1.1.0) still verifies at /verify.