AI murder mystery games that can’t cheat

Can ChatGPT or Claude run a murder mystery?

Yes, and it will run one happily. That's the problem. Ask ChatGPT, Claude, or Gemini for a whodunit and you'll get atmosphere, suspects, and a confident reveal inside the hour; what you won't get is a case that was solvable. Load a sealed one instead: Blackfile №1, a locked-room vanishing, is free, runs 60 to 90 minutes, and its answer was fixed before the file went public.

What’s wrong with most AI murder mysteries?

Most AI mysteries decide the murderer at the end, not the start. Ask a chatbot to run a whodunit and it improvises clues as it goes, then, when you accuse someone, quietly picks whichever culprit best fits the story it happened to tell, or simply agrees with you to be agreeable. There was never a fixed answer to solve towards, so "solving" it is theatre. Nothing stops the model from retrofitting motive, planting a last-second clue, or swapping the killer to make the ending land, and none of that shows on the page.

Is there a murder mystery prompt for ChatGPT that works?

Not one that asks the assistant to invent the case and keep the culprit to itself. If the solution is never written down, nothing holds it still, and the reveal is whatever fits by the end. A murder mystery prompt that works carries the case with it: the suspects, the timeline and every clue fixed on the page before your first question, and the solution sealed in a separate file whose fingerprint is published before you start, so the assistant unseals an answer it could not have changed. That is what each Blackfile card is: a ready-made mystery prompt you open in ChatGPT or Claude with one tap, or paste whole into an assistant that can't open links. Blackfile №1 is the place to start.

How does a sealed solution stop the AI from cheating?

A sealed solution commits the answer in advance with a cryptographic hash, so it's fixed before the first clue is dealt. The culprit, method, and motive are written out and run through SHA-256, a function that turns any text into a fixed 64-character fingerprint, and that fingerprint is published where you can see it before you start. From then on the AI is playing against a solution it can't rewrite. If it tried to change the killer, the real answer would no longer match the fingerprint it already committed to, and the verify page would come back No match.

You'll find the start of that fingerprint on the case's own page, on the seal line beside the link to the verify page, and all of it prefilled when you pick the game on the verify page. The case's manifest file, /g/blackfile-1/MANIFEST.json, carries all 64 characters if you'd rather read them yourself.

How do you verify the seal yourself?

You verify by pasting the reveal into the verify page. The case's reveal file holds the sealed SOLUTION, which can run to many lines, and a random SALT, and the published commitment is the SHA-256 of those two joined by a newline. Paste the whole file into the verify page, pick the game to prefill the published hash, and your browser recomputes it.

1. Finish the case. Make your accusation; the game master reads out the sealed answer.
2. Get the reveal file. Ask for it verbatim, or open its address yourself: it is in the game's play prompt.
3. Paste the whole file. All of it, SOLUTION and SALT included, goes into the verify page; pick the game to prefill the commitment.
4. Check the seal. Your browser computes the SHA-256, and it matches or it doesn't.

The property that makes this work: change a single character of the answer and the SHA-256 hash changes completely and unpredictably. So a matching hash is proof that the reveal is byte-for-byte the answer that was sealed, with nothing edited to fit your guess. No sitemap lists the reveal file and search engines are told not to index it, so the ending never turns up in a search result. Its address is in the game's play prompt and in the card, so the game master can find it. Some cases open it when you accuse; the ones that keep their whole script there read it at load and keep it from you until the end. The methodology explains exactly what gets committed and why the scheme holds.

What the seal doesn’t prove

Be precise about what a hash buys you. It fixes the ending; it can't audit the middle. Nothing in the maths stops a game master from inventing a clue in scene four or leaning on you when you're circling the wrong suspect. That job belongs to the card, not the crypto: a clue is disclosed only when you take the exact action keyed to it, the card forbids evidence that isn't written in it, and the game master is barred from ranking or hinting at suspects. The seal is the part you can check afterwards; the rules are the part that keeps the case fair while you're in it.

Nor is the seal a vault. The card your game master reads is public and carries the reveal file's address, so a determined reader can open the answer early, the way you can read the last page of a novel first. Spoiling yourself was always on the table; no scheme takes that away, and one that hid the card would cost you the right to inspect it. What the seal buys is the part you can't police yourself: the AI can't move the answer once it has seen your accusation.

Which games use sealed solutions?

Every case on the sealed mystery shelf ships with a committed solution, hashed and published before release. Not every one is a murder: Blackfile №1 (locked-room mystery) and Under Caution (interrogation game).

• Blackfile №1: The Vanishing at Larchmont Hall — A 1925 country-house whodunit: an heiress bolts a glass conservatory from the inside and is not there when the doors come down.
• Blackfile №2: A Death at the Wyke Hydro — A snowbound 1929 poisoning whodunit: six suspects, one cup of hot milk, and an answer sealed before you start.
• Blackfile №3: The Long Exposure at Harrowfell — A 1953 observatory whodunit: a death on the dome stair, six suspects who cannot leave the fell, twenty moves, and a culprit sealed before you start.
• The Third Carol: A Christmas Murder Mystery Game — A Christmas murder mystery game for one to eight at one screen: six suspects, four rounds of evidence, and every line sealed before play.
• Under Caution: An Interrogation Game — An AI detective game for one: question three suspects and break their stories against the evidence, ruled by a list sealed before play.
• No Ghost Speaks: A Halloween Murder Mystery Game — A Halloween murder mystery game for 12 and over, one to eight at one screen: five suspects, twelve exhibits, every answer and lie sealed before play.

Across the catalogue, 24 of the 39 games seal an answer this way, from hangman to the quiz nights.

Blackfile №1 is the worked example: the Larchmont Hall answer was sealed before anyone played it, so you investigate, accuse, and then verify that the answer never moved. A case doesn't have to run in one sitting. The game master offers a save whenever you pause, and you can pick the investigation back up in a fresh chat, or even in a different assistant.

The same seal runs under the social deduction shelf: in Ambervale and Dead Air the hidden culprit (Ambervale's liar, Dead Air's saboteur) is committed before the first round, so the AI can't pick one once it sees who you suspect. Browse the rest on the sealed mystery shelf; on every case there, the ending is something you can check.