How We Publish Connections Answers Every Day (Behind the Scenes)

Our daily Connections archive isn't manually written. It's a fully automated pipeline. Here's how.

Step 1: The trigger

A cron job runs every day at 0:30 ET (30 minutes after NYT releases the new puzzle). The script lives on a small always-on server.

Step 2: Fetch the puzzle

We retrieve the daily puzzle data from NYT's publicly accessible game endpoint. The data comes back as JSON with 4 groups, 4 words each, plus difficulty colors.

Step 3: Write to disk

The puzzle is saved as a JSON file in the format data/puzzles/YYYY-MM-DD.json. The naming convention is what makes the archive sortable.

Step 4: Commit and push

The script commits the new JSON file to the GitHub repo. Vercel detects the push within seconds and starts a new deployment.

Step 5: Deploy

Vercel rebuilds the static archive pages — Next.js's generateStaticParams automatically picks up the new date and generates the page. About 45 seconds later, the new puzzle is live on the site.

End-to-end: 1 minute from NYT release

So when you search "connections answer today" at 12:05 AM ET, our page is already indexed by Google and ready to help.