Show HN: Dense NYT style crossword puzzle generation with Rust and Claude (github.com)

A new end-to-end pipeline for generating dense, NYT-style crossword puzzles has been introduced, combining Rust, a constraint-solver fill engine, and Claude Opus 4.7 for AI-generated clues. The pipeline allows users to create puzzles of varying difficulty, from easy to expert, and provides optional explanations for each answer. The tool can be accessed at wordfuzz.com, enabling users to play puzzles generated on their local systems without sending data online. This development is significant for the AI/ML community as it showcases the integration of different programming languages and AI models to solve complex problems in creative fields like puzzle creation. The use of Rust for grid generation ensures efficiency and quality, while the integration of Claude for clue writing and quality assessment introduces a novel way to enhance user engagement through tiered difficulties and detailed explanations. Developers can leverage this pipeline's modular structure, which allows for flexible puzzle configurations and high-quality output while also allowing improvement through user feedback on puzzle clues. Overall, this project represents a blend of AI capabilities and traditional game design, opening doors for further innovation in automated content creation.