For content marketers

Content marketing tools that measure the SERP instead of guessing at it

Decide what to publish from competitor keyword data, write against targets taken from the pages currently ranking, then find out whether search engines and AI assistants actually cite the result.

Ideas from real keyword dataTargets from the live SERPThree live scoresCited or not, per engine

Content marketing has a measurement problem at both ends. At the start, the brief is usually somebody's judgement about what to write, dressed up in a template. At the end, the report is sessions and time on page, which tell you a post was read but not whether it did the job it was commissioned for.

Both ends can be measured instead. What to write can come from the keywords your competitors rank for and you do not, which is a record of decisions somebody else already tested. How to write it can come from the pages currently ranking for that term, which describe what the query actually rewards rather than what a generic checklist assumes.

And whether it worked has a harder test than traffic now that a share of research ends inside an AI answer: was the page cited, and if a competitor was cited instead, which of their pages did the engine read?

Plan

A content plan built from keywords someone already tested

Gap keywords come from DataForSEO, which means the terms your competitors rank for and you do not. A model is used for exactly one job, deciding whether a keyword belongs to your niche, because a large competitor with a large blog ranks for a great deal that has nothing to do with your business. It is never asked to invent a keyword, and anything it returns that was not in the source list is discarded.

  • Competitor gap keywords, not model-generated suggestions
  • Volume and difficulty re-read from the source on display
  • Grouped into page-sized clusters with a primary and supporting terms
  • Anything you already rank for never enters the list
82
A content plan built f score
Brief

Briefs a writer can actually work from

The outline builder turns the keyword analysis into a section-by-section plan: the headings, the terms that belong under each, and the length band the ranking pages sit in. A writer opening that document already knows what the piece has to contain, which removes the round of edits that exists only because the brief was vague.

  • Headings and supporting terms planned before the first sentence
  • Word, heading, paragraph and image targets from competitor pages
  • Every term with a min and max range rather than one number
  • Hand the brief to a writer or write it in the same document
82
Briefs a writer can ac score
Write

Three scores that update while the draft is still soft

Content Score, SEO Score and AI Search Score compute in the browser as you type, covering term coverage, structure, readability and the signals that make a passage easy for an assistant to quote. Feedback that arrives on a keystroke changes the sentence you are writing. Feedback that arrives after a re-analysis becomes a report you read once and file.

  • Live scoring, not a queued re-analysis
  • Import a published URL and score what you already have
  • AI detection with a humanizer for the flagged passages
  • Internal link suggestions from your own sitemap
82
Three scores that upda score
Measure

Whether the work got cited, not just clicked

Once a piece is live, the honest question is whether it changed anything. Search Console data shows the queries and positions the page picked up. The visibility check asks five answer engines the questions your buyers ask and records whether you were named, with sentiment and the list of URLs each engine actually cited on the way there.

  • Real queries, positions and impressions per page
  • Named or not across five answer engines
  • The sources each engine cited, so a miss is diagnosable
  • Runs stored rather than overwritten, so movement is readable
82
Whether the work got c score

How it works

01

Find the gap

Competitor keywords you do not rank for, clustered.

02

Brief it

Headings, terms and length from the live SERP.

03

Write and check

Three scores, detector, links, then publish.

04

Read the result

Search Console positions and answer-engine citations.

The problem with content tools that score against a rubric

Most content optimization tools grade a draft against a fixed list of factors written down once and applied to every keyword in every market. A rubric cannot know that a comparison query rewards 2,400 words and eleven subheadings while the product query beside it ranks 700-word pages, so it guesses, and you write toward a number nobody can trace back to anything.

Measuring the live SERP instead means the target is a description of the pages you are trying to displace, for that query, in that market, on the day you write. When the results change, the targets change, because they were never anything other than a summary of the results.

  • Term ranges taken from the pages currently ranking
  • Structure targets from the same set, not from an average of the web
  • Market and language chosen per query, so a UK piece is measured on UK results
  • Every number traceable to pages you can open and read

Where AI assistants change the job

A growing share of research now ends inside an answer rather than on a results page. Someone asks an assistant which tool to use, gets three names and a short description of each, and clicks one. If your content is not among the sources those three came from, the comparison you might have won never happened.

This is not the same measurement as ranking, and it cannot be derived from a position. An engine composing an answer is choosing what to say and what to cite, and pages ranking fourth get cited while the top result goes unmentioned. Treating an AI answer as a proxy for rank means missing the cases where the two disagree, which is most of the interesting ones.

What makes it actionable is the source list. It names the pages the engine read before answering, which turns a bad result into a specific piece of work: this competitor page is being cited for this question, and we do not have its equivalent.

Optimising what you have already published

New articles are the smaller half of most content programmes. The larger half is the archive, some of which is sitting on page two for terms worth having and would move with an afternoon of editing.

Import a published piece by URL and it is scored against the same competitor-derived targets a new draft would get, so the gap between what you published and what currently ranks is visible before you rewrite anything. Pair that with the site crawl and both halves of the diagnosis are covered: which pages are technically held back, and which are simply under-written for the results they are chasing.

The editor also reads your sitemap when you start a new piece, and flags a slug that already exists or a keyword close enough to an existing page to compete with it. Cannibalisation is easier to prevent at the brief stage than to unpick after both pages are indexed.

Reporting on content without overclaiming

Three sources of truth sit in one place here, and they answer different questions. Search Console says what the page actually earned in impressions, clicks and positions. The scores say whether the draft covers what the ranking pages cover. The visibility runs say whether answer engines name you, and which sources they used.

None of them is attribution, and none of them will tell you a piece of content produced revenue. What they do give you is a report that survives a sceptical question, because every number points at something checkable rather than at a model's opinion.

Runs are stored rather than overwritten, so a quarterly review is a matter of opening two dates side by side. That is a duller kind of reporting than a dashboard with a rising line, and it is considerably harder to argue with.

When the gap list is longer than the budget

A gap analysis on a competitor who has published for six years returns more work than any team can fund. The list is not the plan, and treating it as one is how content calendars end up full of pages nobody had a reason to publish beyond the fact that a tool suggested them.

Three filters usually cut it to something fundable. The first is whether you can plausibly compete for the term at all, which difficulty answers roughly and a look at who currently ranks answers properly. The second is whether the query has anything to do with what you sell, since the relevance gate removes the obviously unrelated but cannot tell you that a term is adjacent rather than useful. The third is whether you already have a page that should be ranking for it, in which case the work is a rewrite and not a new piece, and the editor will tell you how far off the existing text is.

That third case is worth looking for deliberately, because it is consistently the cheapest content work available to a site with an archive. Rewriting a page that already has some authority against targets measured from the current results costs an afternoon. Ranking a new page for the same term costs months.

Whatever survives those filters is a queue rather than a calendar. Clusters do not expire, and the map can be re-run later, at which point anything you have since covered has dropped out of it because the source is a list of terms you do not yet rank for.

Fitting it into a team that already has a process

Most content teams already have a calendar and a CMS and are not looking to replace either. This slots in at the two ends that are usually weakest: deciding what to write, and finding out what happened after publishing.

Drafts go out to WordPress, Ghost, Payload or Strapi directly, or export if a review step lives somewhere else. Briefs can be handed to freelance writers, and since seats are included rather than sold, adding a contractor for the length of a project does not change the bill.

What is not here is a content calendar, an approval workflow or a client-facing portal. If those are load-bearing in your process, keep them. The parts worth moving are the brief and the measurement.

Perfect workflow

One platform. One workflow. The full content loop.

01
Find the gaps
Keyword Clustering

Real keyword data from your competitors, filtered and clustered into briefs.

02
Write & optimize
SEO Content Editor

Write to targets measured from the pages already ranking.

03
Fix what's broken
Content Audit

Crawl the domain and work a fix list ordered by impact.

04
Check the answers
AI Visibility Tracker

See whether ChatGPT, Gemini, Perplexity and Claude name you.

Frequently asked questions

How is this different from a general AI writing tool?+

An AI writer produces text. This produces text against a measured target: term ranges, word count, heading and image counts taken from the pages you are competing with, plus scores that show where the draft still falls short. The AI assist is a drafting aid inside that structure rather than the structure itself.

Where do the content ideas come from?+

From DataForSEO, as the keywords your competitors rank for and you do not. A model filters that list for topical relevance and nothing else. It is never asked to think of a keyword, and any keyword it returns that was not in the source list is discarded along with any volume it tried to supply.

Can I use it for content in languages other than English?+

Yes. A query carries a market and a language, and the results measured are the ones for that market, so the terms, length band and structure targets come from that market's ranking pages rather than from a translated US result.

Does it replace our content calendar?+

No. There is no calendar, approval workflow or client portal here. It covers deciding what to write, writing it against real targets, and measuring what happened. Whatever you use to schedule and approve work stays where it is.

How do I know whether AI assistants cite our content?+

A visibility run puts your prompts to ChatGPT, Gemini, Perplexity, Claude and Google's AI Overview, records whether your brand was named, and returns the URLs each engine cited. Being named and ranking are different measurements, and they disagree often enough that one cannot stand in for the other.

Why do visibility results change between runs?+

Because the answers change. These models search the live web and are non-deterministic, so the same prompt can name you today and skip you tomorrow. That variance is what is being measured, which is why runs are stored and read as a trend rather than trusted as a single score.

Can freelance writers work in it?+

Yes, and it costs nothing extra since seats are included on every tier rather than sold. You can hand over a brief that already carries its headings, terms and length targets, or share an editable link for a review step.

Does it check content for AI detection and plagiarism?+

Both run in the editor before publishing. The AI detector returns a probability rather than a verdict, and the humanizer rewrites the passages it flags. No detector on the market is conclusive, ours included, so read a high score as a prompt to edit rather than as proof of anything.

Your brand deserves to be seen, everywhere.

AI is shaping decisions. If you're not in the answers, you're not even considered. Win the next wave of search.

Get started now