FOR SMALL PUZZLE CREATORS
A clue can sound right while leaving two valid answers. NALOO offers a narrowly scoped, AI-assisted constraint check for one English logic-grid or finite code puzzle: up to four entities, three categories and twelve clues. We first confirm that the rules can be represented exactly. No payment is requested before scope confirmation.
WHAT YOU RECEIVE — US $19
A plain-text report containing the explicit assumptions, modeled clues, number of valid solutions, any counterexample found, and one suggested clue repair if needed. The Python checking script and one clarification pass are included. This is not human playtesting, a difficulty rating, a proof about ambiguous language, or a guarantee of commercial success. We do not accept live contests, exams, private customer data or puzzles you lack permission to share.
A REAL EXAMPLE FROM OUR OWN PROTOTYPE
The first Night Manifest draft had two valid assignments: Ava could carry either the book or the map. Adding “Ava did not carry the map” reduced the enumerated result to one. This is our own development example, not a client result. The corrected free puzzle and explained answer are downloadable at https://hongki69.itch.io/the-last-label .
HOW TO REQUEST A CHECK
Contact NALOO on Reddit at https://www.reddit.com/user/msjdg1469/ with “Puzzle check”, the category sizes and your deadline. Initially send only a public puzzle link or a non-sensitive summary. We confirm scope and timing, show a short finding excerpt, then offer the completed report through Payhip if you want to purchase it. No account access is needed. This page does not take payment.
AI-assisted modeling can be wrong: the report explicitly maps every clue to its constraint and states what was not modeled. Keep your own editorial review. NALOO is operated through Little Date Studio.