Symptom guide / Publishing and display
A blank screen after you publish
The published address opens, but the screen stays empty even though the editor preview had content.
Start with the exact address a visitor uses, including any path after the domain. A blank page alone cannot tell you which part failed. This check separates delivery of the page from delivery of its scripts, then looks for a failure while those scripts run. Keep the preview open for comparison, but record evidence from the live tab.
01 / Procedure
The two-minute check.
- Open the live address in Chrome.
Right-click the empty page, choose Inspect, and select Network. Reload once with that panel open. Start with the row whose Type is document; read its Status and open Response to see whether the host actually returned your app’s HTML.
- Check the files needed to draw it.
Click the JS filter. Open a failed script row and read Headers and Response. Record its request URL and status. A response containing an HTML page where JavaScript was expected is useful evidence of a routing or asset-path problem.
- Read the first console failure.
Select Console, reload again, and record the first error plus the file and line shown beside it. Use the actual message from your page. Later messages may follow from that first failure; do not invent an error to give the builder.
- Compare a fresh visit.
Open the same live address in an incognito window. If it draws there, record that difference before clearing anything in your normal browser. If both fail, compare the failing file URL with the published build’s files.
02 / Evidence
Read the result.
WORKSThe live page draws in a fresh visit and the required script requests complete.
BROKENThe document or required script fails, or startup stops at a reproducible console error.
03 / Repair
Causes and fixes.
Ranked in the order to investigate, not by claimed frequency.
- The document or a script is missing.
Correct the host’s published output directory or the app’s asset path using the failed URL as evidence. Verify that the replacement deployment actually contains the requested file before retesting the live address.
- JavaScript starts and then fails.
Give the first console failure and its file location to the person changing the code. Repair that failing startup path, then reload with Console open and confirm the page renders.
- The browser has an older copy.
Use the cache guide to compare the deployment address and browser cache before changing application code. A fresh window working is a clue, not proof that every visitor is fixed.
04 / Limits
Edge cases.
- If only a bookmarked inner page fails, compare it with the home address and then navigate there through the app. That isolates direct-route handling from general startup.
- A page can render text that matches its background. Check the Elements panel before assuming every visually empty page is a JavaScript crash.
When to stop and hand it over
Stop after capturing the document status, first failed script and first console error if the fix requires editing unfamiliar startup code. Send the exact URL and a screenshot with tokens removed. Stop sooner if a response reveals private information.
05 / Questions
Before the next attempt.
Does a successful publish prove the screen works?
No. Open the published page and repeat the check. A finished build and a usable browser screen are separate observations.
Should I paste the whole console into a public chat?
Copy only the relevant failure after removing private URLs, account data and tokens. Keep the original privately for the repair.