FAQ

WordProxi is a semantic word guessing game where you attempt to find a secret target word based on its contextual similarity to your guesses. Unlike traditional spelling-based games, WordProxi uses vector-space linguistics to measure how "hot" or "cold" your guess is relative to the target.

Type any common English word into the terminal and submit it. The system returns a similarity score so you can keep exploring nearby ideas until you find the hidden word.

Similarity is calculated using cosine similarity in a high-dimensional word embedding model. It doesn't look at letters or prefixes, but rather how often words appear in similar contexts across billions of documents. 'Doctor' and 'Hospital' have high similarity, while 'Doctor' and 'Donut' have very low similarity.

While both are daily word games, the mechanics are fundamentally different. Wordle is a combinatorial game about letter placement. WordProxi is a semantic game about conceptual proximity. You won't find clues in the letters; you'll find them in the meaning.

The scanning array accepts all standard English lemmas. Pluralization is recognized, but be aware that 'Dog' and 'Dogs' occupy slightly different points in the vector space, though they will usually both be very 'hot' if the target is related to canines.