Search Console Query Reader

A free tool for founders and content leads asking what people already search for when they find them. Paste a Google Search Console Queries export and the reader shows impressions and clicks by intent, groups the queries into topics, names the topics you already appear for on page two or worse, finds queries split across two of your own pages, and gives the branded share of your clicks. It runs in your browser; nothing you paste is sent anywhere.

How it reads the export

It runs the engine of the AAJ search-demand-reader agent skill, unchanged. Each query's intent is read from its words: your brand words make it branded, a rival's name makes it rival-brand, words such as demo, trial or pricing make it transactional, best, vs, alternative or software make it commercial, and how, what, guide or template make it informational. Anything else is unqualified, which means the words did not say. Queries are grouped into topics by their most common word, after light stemming.

The engine's rules

• A topic needs 200 or more impressions to count as demand.
• A topic whose average position, weighted by impressions, is worse than 10 is one you already appear for on page two or worse.
• A query is split across your pages when two or more of them each take 20% or more of its impressions. This needs a page column, which the standard Queries file does not have.

These cut-offs are the engine's rules, stated so you can argue with them. They are not industry benchmarks.

What it can't see

• Impressions count the times your pages were shown. They are not search volume, and the reader never estimates one.
• Searches where you never appeared are not in the export. The reader cannot size a market or fill that gap.
• Branded share depends on the brand words you list. Leave one out and its searches count as non-branded demand.

Why I built it

Founders ask me what people search for, and the usual answer is a keyword tool's estimate. Search Console already holds a better answer for anything you show up for: a real count of the times Google chose you. I wanted a reader that stays inside that count and says plainly what it cannot see.

The same engine runs as an agent skill, search-demand-reader. To see which pages are losing clicks over time, use the Content Decay Finder. To brief the page that closes a gap, use the Content Brief Builder.

Run this in your agent

The free AAJ agent skill Search Demand Reader does this job inside a coding agent, on your own data. Install it with npx skills add sarojkjha/aaj-marketing-skills --skill search-demand-reader.

Using Claude or ChatGPT instead? Add the AAJ engines as a custom connector, with no install: https://aaj-engines-mcp.sarojkjha10.workers.dev/mcp