I like knowledge games, but I often feel they stop at “right or wrong.” Webwoven started with a different question: what if finding the connection was the actual game?

A round gives you two real things, for example Lionel Messi and the Tour de France. Your job is to find a path between them through real, named relationships. Every move tells you why the connection exists, and you can inspect the fact and its source before choosing it. Sometimes the obvious direction works. Sometimes you end up somewhere completely unexpected and have to go back.

There are three ways to play. Single player is for exploring at your own pace. The Daily challenge gives everyone the same start and target. Multiplayer lets two to four people race through the same graph in a shared Lobby. It works in the browser, on desktop and mobile, and does not need an account.

I built Webwoven as a solo project during Build Week with Codex and GPT-5.6 Sol. I brought the original idea, rules, visual direction and a lot of fairly specific opinions about how the game should feel. Codex helped turn that into the web app, backend, data pipeline, tests, documentation and deployment setup.

The process was not one big prompt that produced a finished game. I kept testing it on my laptop and phone, and other people tried it too. That uncovered repeated routes, confusing labels, bad mobile layouts, false dead ends and several Multiplayer timing problems. I brought those concrete failures back to Codex, checked the fixes and usually found another smaller problem afterwards. A large part of the final game came from that loop.

GPT-5.6 and Codex helped build the product, but they do not play the game for you. During a round there are no calls to an AI service, Wikidata or Wikipedia. Webwoven runs from a checked local atlas containing almost 4k playable entities and over 20k named connections. The published start and target pairs are validated, but the path you take between them is still yours.

It is already live and playable, but I do not consider it “done forever.” I want to keep watching where new players hesitate, improve route variety and make the first round easier to understand without explaining away the surprise.

Built With

Share this project:

Updates