Back to browse
GitHub Repository

Dense NYT-style crossword puzzle pipeline: Rust fill-engine + Claude clue writer/QA/explain process

1 starsRust

Dense NYT style crossword puzzle generation with Rust and Claude

by ekorbia·Jun 8, 2026·3 points·0 comments

AI Analysis

●●SolidBig BrainNiche Gem

Rust constraint solver generates dense grids while Claude writes and QA's clues automatically.

Strengths
  • Rust constraint-solver fill engine handles dense grid generation as a genuine CSP challenge
  • Multi-tier AI clue writing with adversarial QA review loop catches and fixes flagged clues
  • Client-side wordfuzz.com renderer lets you test generated puzzles without server uploads
Weaknesses
  • Claude Opus API dependency means the clue-writing pipeline isn't free or offline
  • Crossword generation is a niche use case with limited developer audience appeal
Category
Target Audience

Crossword enthusiasts and puzzle developers

Similar To

Crossword Labs · Puzzle Maker · Cruciverb

Post Description

I’m a daily New York Times crossword player and I’ve built xword-pipeline with the goal of creating my own dense New York Times style crossword puzzles.

The pipeline consists of a Rust grid fill-engine and a Claude based clue-writer. The fill-engine generates candidate grid skeletons (default is 200) and fills them with words from the Crossword Nexus Collaborative Word List. The clue-writer uses Claude to generate multi-tiered clues (easy, medium, hard), explanations, and QA review the generated puzzles. The clue-writer explanation and/or QA steps can be skipped or the clue-writer can be skipped entirely if you want to write your own clues.

I’ve also created a crossword puzzle web app that can be used to test puzzles generated by the pipeline (https://wordfuzz.com/). The puzzles in the archive have all been generated with the pipeline. QA reviews did show issues with some of these puzzles but they should be solvable and I’ve left them as is so the quality of the pipeline can be accurately judged. You can run the pipeline yourself and use the test page ( https://wordfuzz.com/test) to test the puzzles you generate. The test page uses LocalStorage so nothing is uploaded to the server.

Note: The clue-writer requires an Anthropic API key and costs me an average of around $0.50 to fill a 15x15 grid with multiple-tiered clues and explanations. If you’d like to give the pipeline a try, run a single puzzle first and verify the cost is acceptable to you before generating multiple puzzles.

Any feedback on the pipeline would be appreciated. I would be great to hear if you’ve had success generating fun, playable crosswords or if you have any suggestions for improving the pipeline. Thanks!

Similar Projects

Gaming●●Solid

I Made a Math Crossword

The project grafts hexadecimal arithmetic onto a tiled crossword mechanic — you pick squares or shapes that represent a target hex number, with difficulty toggles, a timer, and an on-demand hint (press H). The UI is clean and immediately playable, but the core idea is a clever niche twist rather than a broad innovation; it could use better onboarding, scoring explanation, and shareable hooks to make it stick.

Niche GemCozy
keepamovin
203mo ago