RankRecon
Turn the Google SERP into actionable data
RankRecon is a free browser extension for Chrome and Firefox. You run it on a Google SERP and it reads the results page and every site ranking in it, entirely on your own machine, then gives you back a dataset you can save and an analysis of what's winning the SERP.
Where it came from
I run a small agency doing SEO and web work for SMBs. When a client comes to me, the first thing I want to know is how they're ranking. I use Ahrefs and it's brilliant - backlinks and DR quickly tell me how the client is doing. But I can't get the full story on the rest without going and looking at the actual SERP - what is Google putting on the page, where is my client ranking, who is winning here and why.
That stuff takes time, and I end up going through multiple SERPs, ten tabs open for each one and trying to hold it all in my head for long enough to get a read on what's going on. And I do that over and over.
So I built a thing to save me from doing that. I point it at a SERP, get back what I need as data, saved out with a timestamp. The saved file turned out to matter as much as the analysis - it's a record of where a client and their competitors stood the day we started, and I can run it again in three months and see what actually moved.
What it does
Extracts the SERP: organic results, the AI Overview and who's cited in it, and the Local Pack. You can walk up to 10 pages of results and it stitches them into one dataset.
Measures every page that ranks: word count, the actual H2 and H3 text rather than just counts, internal and external links, images, schema markup, and a couple of dozen other structural facts.
Gives a read: intent per page and for the SERP overall, with results clustered by shape so you can see which content pattern is holding the top positions.
Exports to CSV, Excel, JSON, Markdown, HTML or plain text.
Copy for AI - one click gives a version of the SERP data you can paste into your preferred AI model for further analysis.
How it's built
It all runs in your browser. No server, no account, and nothing you search for leaves your machine.
When you run it, it reads the SERP first. As well as the organic results, it maps the other blocks Google is showing - ads, videos, People Also Ask - so you get the shape of the whole page, not just the ten links.
Next it visits the sites. Each result is loaded in a background tab and measured. Hidden elements, pop-ups and cookie banners get filtered out first, because a consent dialog with 200 headings in it will otherwise wreck your numbers.
The report is then computed from those measurements. Each page gets an intent classification from a weighted read of its schema, title, URL and structure, and the same signals decide the overall intent of the SERP. Pages are clustered by shape - length, heading density, link density - into content patterns, with whichever pattern holds the top positions flagged as what's currently winning. Last, it cross-references which AI Overview citations also rank organically, and which Local Pack businesses appear in the organic results.
It's built on the 80/20 rule, so it's deliberately not exhaustive and it will be wrong sometimes. What it gives you is a quick read on a SERP that would otherwise cost you an hour.
Where it's at
Free, with no catch. It's early, and right now it's a tool shaped around my own habits. What I don't know yet is whether the core of it is useful to anyone else. If you do SEO and you run it on a SERP you already know well, I'd really like to hear where it's wrong or what's missing.
Comment