SEOLinkMap Master Workflow

The workflow runs in one order: crawl the site, fix what the audit finds, then research one keyword at a time. Each order comes back as work to do on named pages.

Keep that keyword the same across the tools. Competitor Research and Backlink Prospector are both driven by the keyword you pick, and AI Visibility is configured by prompt, so write the prompt around the same subject. Finish one keyword, then batch the next set behind a new one.

If you would rather be told which step you are on than work it out, what do I do next reads your account and names it.

Crawl And Audit First

Add your domain. The crawler builds the LinkMap and the audit together, so you get the structure of the site and the list of what is broken in the same pass.

Fix audit items however you work -- by hand in your CMS, or through the private MCP from your editor. How To Do SEO With Claude Code covers that loop.

Either way, the MCP will re-measure a single page on request, so you can confirm an item is resolved on that page instead of assuming it.

Research One Keyword

Competitor Research takes one keyword and takes apart the SERP behind it: the pages outranking you, measured on content, speed and structure. This is the keyword the rest of the sequence lines up behind.

The output is alignment work on your target page -- what the winning pages have in common, and what you can leave alone. Findings land in four buckets: OPPORTUNITY, CRITICAL, AVOID and GAMBLE.

With the private MCP connected, the LLM reads what SEOLinkMap measured rather than guessing from training data. Your SERP, your pages, your numbers.

Build AI Visibility Over Time

AI Visibility measures which pages Google and the public AI chats already pick when someone asks your question, then points at what those pages have that yours does not. It is configured by prompt rather than by keyword, so phrase the question a buyer would actually ask about the keyword you are already working.

It is the one tool here that needs time rather than a single order, and the payoff usually arrives before a ranking does. Being named in an AI answer does not wait for you to reach the top of a results page. The methodology is published in full.

Over the MCP, you can ask it to bring a target page in line with the ones already being chosen.

Prospect Backlinks Last

Leave outreach until the site is healthy and the critical audit items are gone. Links pointed at a broken page are wasted. Prospect on the keyword you researched, so the sites you approach are the ones already writing about it.

Backlink Prospector spends hours of research per keyword finding sites worth approaching, and hands back a draft pitch written by an LLM. Edit it before it goes out -- it is there to kill the blank page, not to be sent as written.

Over the MCP it will sharpen the draft and keep track of who you have contacted. Add a Gmail MCP alongside it and the whole outreach loop runs from one chat, on real data, with millions of tokens reduced to a few tens of thousands.

That closes the keyword. The next one starts the same way: fresh competitor research, a prompt written around it, and its own prospect list.

Plain English With Evidence

None of this needs the MCP. Everything is in the site when you log in, and the MCP is a second door rather than a requirement.

Wherever a number drives a recommendation, the sentence you read is built from that number. You get what to do and what to skip, instead of a page of statistics you have to interpret yourself.

Then click Evidence. The charts, tables and correlations the sentence was built from come back in full. The plain wording is a summary of the math, never a replacement for it.

What An Order Buys

Tools that index the whole web can hand you unlimited keywords and a score out of a hundred. That breadth is real, and SEOLinkMap does not match it at this time.

One keyword here buys something different: days or weeks of specific work on named pages, with the reasoning attached. More keywords cost more over time, because each one is another body of work rather than another row in a table.

A score out of a hundred tells you where you stand. It does not tell you what to do on Tuesday.

Supercharge with Private MCP

The private MCP server connects your own SEO data to Claude or another LLM, so you can ask for it in words instead of reading it off a dashboard.

Set up agentic SEO workflows that measure a page against the competitor research already run for that keyword. For new content, give the LLM your targets and let it check the draft against the same findings before you publish.

The questions that pay off are the ones a dashboard cannot answer in one view. Which service page has the most to gain from the current SERP data. Which pages moved after the last round of internal linking. Whether the post you just wrote matches what the top-ranking pages are doing.

Connect to SEOLinkMap's MCP server once, and every step above is available as conversation as well as dashboard.