JevMade Sign in
← Back to experiments

Apps & data pipelines

xword

A crossword agent where Python manages the grid and crossings while Jev ranks short wordlist candidates for each clue.

Bookmark: xword Keep this in your collection.
Leave a noteWhat would you try with this? : xword

Only you can see your notes.

Source screenshot of xword
SOURCE SCREENSHOTFull screenshot ↗

What it does

Only clues with tight crossing patterns reach Jev. A separate language model handles regions where the wordlist cannot supply plausible candidates.

How you can use it

For a crossword helper, ask a developer to follow this project's rule that letters must agree where words cross. Once two letters are known, AI ranks words from a short list. Another AI fills the remaining areas, then a web search tackles blanks. The code checks the crossings instead of trusting every word the AI suggests.

Maker-reported (not independently measured by JevMade): Race on the same puzzle ids: 3×3 solved in 6.0–6.8 s with 0 Token Factory tokens against the LangGraph baseline's 12.1 s / 3,116 tokens (author-reported) · 64×64: 511 s and 3,598 Token Factory tokens (~$3.85) against 2,009 s / 17.7M tokens (~$30.90) (author-reported) · The earlier LangGraph agent took 98 turns, 996 s and 2.44M tokens on a 13×13 (author-reported)

Primitives
choice, noul
Platform
Python
Added
Project created
GitHub stars
0 (snapshot, not live)

Keep this for later

Sign in to bookmark experiments, guides and videos, and keep notes only you can see.

Continue to sign in

We’ll bring you back to this listing.