Built from the SERP that exists today

Before generating anything, the tool queries Google for your keyword and pulls the top 10 organic results with their positions, titles, snippets and domains, plus the People Also Ask questions and related searches. Those signals go into the prompt as evidence, and the brief is written to beat those specific pages. Every result carries a badge showing whether the run used live SERP data or fell back to inference, so you always know which one you are reading. Optional audience and domain fields sharpen the framing.

What lands in the finished brief

The output is structured rather than prose, so a writer can work straight down it. You get the intent classification, a description of who the piece is for, a recommended word count, and an outline of 5 to 7 H2 sections with talking points under each. Below that sit the entities to cover, the questions to answer, internal link ideas, and the competitor pages that were analysed, linked so you can open them. The outline exports to CSV or Excel as section and point rows.

• ›Search intent and target audience
• ›Meta title and meta description drafts
• ›5 to 7 H2 sections, each with talking points
• ›Entities, questions grounded in People Also Ask, internal link ideas
• ›Up to 8 competitor URLs that were actually read

One keyword, or twenty-five

Single mode is for the brief you need right now. Bulk mode takes a pasted list, a spreadsheet upload or a Google Sheet, runs up to 25 keywords through the same pipeline, and returns a table of intent, word count and meta title per keyword that exports as one file. That is usually enough to cover a quarter of publishing in a single sitting, which is the point: the bottleneck in content operations is rarely the writing, it is deciding what to write and against whom.