Search
Google Search Console
Queries, impressions, positions and index coverage, per property.
Feature
Map the topic before you write, using data you already own
What to write next, and why · Runs in Claude Code
What it does
Start from the queries your site already gets impressions for, rather than from a blank keyword tool — then widen out to seed research when your own data runs out. The system extracts the phrases that keep repeating, weights them by real impressions, classifies the intent behind each one, and maps the topic into a plan: what to write, in what order, and which pages each piece should link to.
$ /ngram-analyzer --site yoursite.com --source gsc
84 phrases, weighted by impressions
$ /topical-map --topic "technical seo"
1 pillar · 14 supporting articles · link plan
§ Scope
7 Skills involved
2 Specialist agents
4 Data sources read
§ Limits
What this does not do
Stated plainly, so you do not find out later
Search volume comes from DataForSEO, which is a third-party estimate — the same class of number every keyword tool sells, and no more accurate here than there. The exact numbers are the ones from your own property: impressions, clicks and position.
§ Skills
7 of them
The commands behind it
Each one is a single command against real data. They compose: the output of one is the input of the next.
Research keywords and keep the ones worth tracking
Explore a seed topic for volume, difficulty and intent, then save what you want and push it into tracking.
You get back A saved keyword library, priced before you spend.
Map a topic before you write about it
Pillar, clusters and articles, with everything you have not covered marked.
You get back A topical map, ready for the pipeline.
Decide what to write next
Turn your opportunity data into an ordered queue of topics.
You get back A prioritised topic queue.
See what people actually want
Sort your queries into researching, comparing and buying — and watch the mix shift.
You get back Intent mix, compared to last period.
Find the phrases that keep repeating
Pull recurring two- and three-word phrases out of your queries, an article, or a live results page.
You get back Ranked phrases, weighted by impressions.
Build a local growth roadmap
Sequenced work across Business Profile, reviews, citations and content.
You get back An owner-facing roadmap.
Plan service-and-city landing pages
The whole local structure, from pillar down to each service in each city.
You get back A briefed set of local pages.
§ Agents
2 on it
Who does the work
An agent holds a role rather than a command: it decides which skills to call, in what order, and what the results mean together.
§ Sources
SERP data
DataForSEO
Keyword volumes, competitor rankings and SERP features.
SERP data
Serper
Live SERP lookups when a check has to be current rather than cached.
Crawl
Firecrawl
Full-site crawls and page extraction, including rendered content.
§ Elsewhere
The rest of what it does
AI Visibility Cited by ChatGPT, or not Backlink Analysis Who links to you, who hurts Rankings & Search Console Your real positions, read properly Technical Site Audit What is quietly broken Content Production Brief to published, with images Competitor Intelligence Their plan, taken apart Client Reporting The monthly deck, written