I built an SEO system for my tools site, but it isn’t allowed to publish anything

I’m building a small tools site called Ziannah with things like random wheels, timers, work and pay calculators, converters, and short helpful articles.

The site is still new, so I know I can’t compete for large keywords yet. At the same time, I don’t want to fill it with thin AI articles or generate hundreds of pages just because a keyword exists. So I built a small system to help me find SEO opportunities, with one strict rule:

It can find and score ideas, but it cannot write or publish anything.

Right now, it:

1. Pulls data from Google Search Console

2. Matches search queries with existing pages

3. Scores pages that may be worth improving

4. Filters out searches that Google can answer directly

5. Stores the evidence and scores in Supabase

6. Runs automatically on Cloud Run

7. Leaves the final decision to me

The first useful opportunity it found was:

• Search query: yes no maybe wheel

• Existing page: /tools/wheel/yes-no-maybe

• Average position: around 6

• Impressions: low, but consistent

• Clicks: 0

Instead of creating another article targeting the same query, I improved the existing tool page. Now I’m leaving it alone for a while to see whether the click-through rate improves.

The next step is adding another source using Keyword Planner exports to find possible new tools and articles. But the same rule will apply: it can suggest opportunities, not automatically turn them into content.

I’m still working out the balance between automation and judgment here. Would you focus on improving pages already getting impressions first, or spend more time finding completely new opportunities?

And for anyone using a similar system, what would make you confident enough that a keyword deserves its own page?

Here’s the site in case anyone wants to see: https://ziannah.com