Preview
What it is.
Ours, and we sell the service it runs inside, so treat the numbers as ours to prove. It takes a target keyword, a client domain and a brand kit, and returns a brief a writer can start from.
The pipeline runs 25 nodes across 4 models. Google Search returns the top 5 results, a Python node drops the URLs that are not content, each surviving page gets scraped and converted to markdown, and Semrush pulls organic keywords in parallel.
Brief writing went from 45 to 60 minutes down to under 10. The reason worth stealing is that the SERP data is pulled fresh each run instead of pasted in by whoever is briefing.
What you get.
- Keyword data with intent classification
- An H1 with alternatives, plus a stated content angle
- A full H2 outline and an FAQ section
- Internal link suggestions and a schema recommendation
- Writer notes, which is the part that saves the follow-up call
How to set it up.
Start from the search results, not from a keyword tool. The top 5 organic results are the input, because they define what the page has to beat.
Filter the URL list before you scrape. A Python step that drops non-content URLs is what keeps navigation pages out of the analysis.
Convert each page to markdown before analysis, so structure survives and the model reads headings as headings.
Pull organic keywords in parallel with the scrape instead of after it. Serial steps are where the 45 to 60 minutes went.
Classify intent explicitly and write it into the brief, because that single line decides the page format.
Give the writer alternatives for the H1 and a stated angle. A single title with no angle gets rewritten by the writer anyway.
Pricing Plans
Not sold as a skill.
It runs inside client engagements, and it needs a Semrush subscription. The published account dates from April 2026.
Use cases
The weekly briefing run
You publish weekly and each brief used to eat most of an hour. Hand the pipeline the target keyword and client domain, and a writer-ready brief comes back in minutes with headline alternatives, a stated angle, a full outline and an FAQ section.
Research that stays current
The old template carried SERP notes pasted in weeks ago. This one scrapes the current top results on every run, so the outline is built against the pages your draft has to beat today.
Skipping the follow-up call
Writers used to ring back with questions the brief didn't answer. The writer notes section carries the angle and intent decisions in prose, so the questions get answered before they're asked.
Blueprint for your pipeline
You're designing a brief generator of your own. The published node order, from filtering scraped URLs through pulling keywords in parallel with the scrape, is most of the architecture handed to you.
Best for
Briefing at volume
The saving repeats on every brief, so the pattern pays off for a team publishing weekly instead of monthly.
Keeping SERP data current
Pulling the top results fresh each run removes the stale-research problem that comes with a template document.
Designing your own brief pipeline
The node count, the model count and the scrape-then-filter order are published, which is most of an architecture.