The Bug That Made Google See an Empty Tools Page
For weeks, Google saw my tools directory as an almost empty page. It listed more than 150 working tools. A visitor could see every one of them. A crawler could not.
Here is the bug, how I found it, and the fix.
How the pages were built
The tools hub, the twelve category pages and the related tools strip on every tool page all read from one data file. A script loaded that file in the browser and drew the grids on the fly. It was tidy: add a tool once and it showed up everywhere.
The catch was the phrase in the browser. The HTML the server sent held an empty container. Everything else appeared only after JavaScript ran.
How I noticed
AdSense rejected the site for low value content. That made no sense for a page with 150 tools on it, so I looked at the page the way a crawler first sees it: the raw HTML, scripts ignored.
The visible tool count read zero. Category pages showed a heading, a short intro and nothing else. To anything that did not run my script, the site was a pile of nearly empty pages.
Search Console agreed. A large share of tool pages sat in Crawled, currently not indexed or Discovered, currently not indexed. With no plain links between sister tools, crawlers had few paths into the site at all.
The fix
I moved the rendering to the server. WordPress now writes the full tool grid into the HTML itself, from the same data. The script only adds extras like filtering and search on top.
- The tools hub ships every tool as a real link in the page source
- Each category page ships its full grid
- Tool pages link to their sister tools in plain HTML
- If JavaScript fails, every page still works
A second bug hiding in the same file
While checking the counts, I found two tools that never appeared anywhere: the background remover and the image to text converter. Both carried a flag in the data file that every script treated as hide me. Search, category grids, site totals and related tools all skipped them silently.
The page count on the site also drifted. The hub said 154, a category said 9 when it held 11, and the real total was 151. Hard coded numbers in several places will always drift. Now the numbers come from the data.
What I would tell anyone building this way
Client side rendering is fine for what only matters after someone arrives, such as a filter or a calculator result. It is risky for anything you want search engines or reviewers to count. If a page exists to list things, put the list in the HTML.
A quick test takes ten seconds. Open the page, view source, and search for a word you can see on screen. If it is not there, a crawler may not see it either.
