Understand every post, entity and link you own, right down to the rel attribute, and what to do about each one.
ContentMapped reads every post on your site, or across all of them, works out what each one actually covers, maps how they connect, and shows you the ground you have left empty. Then it helps with the next post: what to check before it goes live, and which of your existing posts should be linking to it.
The first crawl
It reads every published post on every site you connect, and nothing it does can change what a post says.
Every post placed and grouped into topic clusters, with pillar, sub-pillar and supporting tiers. Size is how central a post is, not how long it is.
The full internal link graph, plus the hygiene problems hiding in it: internal nofollows throttling your own pages, outbound affiliate links missing rel="sponsored", dead links, flagged domains.
Orphans nothing links to. Clusters standing without a pillar. Structured data naming the wrong entity. Entities no post yet owns. None of this comes out of a spreadsheet.
The structure on this map is a real crawled site, anonymised. Only the shape left the database.
The blind spot
You can find out what a competitor ranks for in about four seconds. Finding out how completely your own posts cover a topic, which ones are duplicating each other, and which ones nothing links to takes a spreadsheet and a weekend, and it is out of date by Monday.
So the same three things keep happening. You write a post that overlaps something from 2023 you had forgotten about. A cluster sits there with no pillar holding it together. A post that should be pulling weight has three internal links pointing at it and none of them use the words it wants to rank for.
None of that is a keyword problem. It is a knowing-what-you-have problem.
What it does
Every post, on one site or twenty, indexed and kept current, placed on the map and grouped into clusters. The full internal link graph, with orphans and weakly linked posts surfaced. Change detection, so you can see what moved since last time.
Every entity in your content matched to its Wikidata identifier, held in a registry per site, with salience scoring that tells you how central each entity is to each post. Suspect matches go to a review queue. Your corrections are locked and never overwritten.
Read what each published page's structured data actually claims, and verify every identifier in it. Catch undisclosed affiliate links, internal nofollows, dead internal links, and links to flagged domains. Schema problems come with a corrected, paste-ready block. Every link problem names the exact post and the exact link.
Entity coverage across a cluster shows you which important entities no post yet owns. Only subjects resolved to a real thing are listed, so a name that could mean any of a hundred places is left out rather than ranked. Add ideas to a scored backlog that flags duplication automatically, then export a pre-filled brief for whichever writing tool you use.
The publishing assistant is a mechanical pass over a draft before it goes live: title and description lengths, the slug, paragraph lengths, alt text and captions, image weight, link attributes, affiliate disclosure. No AI and no opinions, just the checks you would make yourself with enough patience. Click any finding to land on that exact sentence in the post. Then, once it is published, which of your existing posts should link to it, and the paragraph in each where the link belongs.
Nineteen crawlers checked one at a time and grouped by what they are for, because "is ChatGPT allowed?" has three different answers. Search engines, the crawlers behind AI answers, the fetch that happens when someone asks an assistant to open your page, and the ones collecting training data. Your robots.txt is compared with what your server actually did, and where those two disagree, that is the finding.
How it works
An application password, and nothing to install. Reading your content needs no plugin at all.
Posts, links, topics, entities, clusters. The first scan covers everything you have connected, then it tracks changes.
Every finding ends in a copy-paste fix or a decision you make. Re-scan when you are done and watch it clear.
On purpose
Most of these are things a competitor would sell as a feature. They are absent on purpose.
Every write is off until you turn it on, and even then it waits: you read what has been staged, and you press push. A tool with write access to two hundred posts is one bad inference away from a very long evening.
There are three of them, each with its own switch, because agreeing to one is not agreeing to the others. Categories and tags. The rel attribute on a link, which is the only one that edits the body of a post, and only that attribute. An entity block, through the plugin. Each is staged for you to read first, snapshotted before it runs, and reverts as a batch with one click. Everything else it finds ends in something you copy, paste and re-scan.
Every AI-assisted call ends at a button you press. If you overrule it, that correction is locked and stays locked.
No made-up entities, no guessed identifiers, no fabricated authors. An identifier is verified against the live source or it is left out. There is no middle setting.
The opportunity scoring is an internal measure of topical authority across your own content. It is not rank data and it is not labelled as if it were. Search Console integration is on the roadmap and will be named as such when it lands.
Where it goes next
These are in motion, and a feature list that hides its gaps is worth nothing, so here they are before they exist.
Pushing article edits into WordPress is not on the roadmap. See above.
Who it is for
Someone running a content site they care about, alone or nearly alone. Ten posts or two thousand, it is the same job: ten posts still deserve structured data that is correct, entities pointing at the right records, clusters forming the way you intended, and internal links using the words the post wants to rank for. There is just less to read.
There will always be more findings than evenings. ContentMapped does not hand you a priority score and tell you what matters most. It gives you the facts to judge that yourself: how many, which posts, and exactly what is wrong. The call stays yours, grounded in something you can open the post and check.
If you run more than one site, it holds all of them and rolls up what each one needs: how many links want reviewing, how many posts have no entity markup, what is missing where. Not connections between sites, each one is mapped on its own, but one place to see how much work is waiting and where.
You will have met words like entity, cluster and structured data. They might already be part of how you work. They might be words you have read a hundred times and never quite pinned down. Either is fine. ContentMapped uses the proper terms rather than inventing friendlier ones, because they are the words you will meet everywhere else and they are worth knowing.
WordPress today, which is what it connects to rather than what it is.
And if you outgrow it and move on to the heavier professional tooling, that is the job done. You will know what you are looking at when you get there.
Questions
Only what you switch on. Three things can be written, each off by default and each with its own switch: categories and tags, the rel attribute on a link, and an entity block through the plugin. Only the rel attribute touches the body of a post, and it changes that one attribute and nothing else. Every write is staged for you to read, snapshotted before it runs, and reverts as a batch. Everything else it finds ends in something you copy, paste, and re-scan.
Not for the analysis. It connects over the WordPress REST API with an application password, and reading your content needs nothing installed. Pushing an entity block back is the one feature with a plugin, and it is optional.
No. It measures what your own content covers and how it links together. It does not report positions or keyword volumes, and it will not pretend to.
A record of every named thing your content talks about, matched to a stable public identifier so that different spellings of the same thing are understood as one thing. It is what makes coverage and gap analysis reliable rather than guesswork.
For some judgements, yes, on your own API keys, and always as an opt-in enhancement rather than the default path. The core analysis runs free without it. Nothing an AI suggests is applied without you clicking.
Not today. WordPress is the first source rather than the only one planned.
Early access
Early access is opening in batches. Leave an email and you will hear when a place is free.