19:36Nico | AI Ranking · Tutorial
How to Perform an SEO Heist Using Claude/Codex
A four-pillar diagnostic — fundamentals, content, backlinks, AI citations — for finding exactly where a competitor is weak enough to beat.
July 23rdThe argument in one line.
Adobe's $1.9 billion acquisition of Semrush prices out functionality that DataForSEO's API plus Claude Opus 5 can reproduce and improve on for the cost of API credits and an existing Claude subscription, because reasoning over the underlying data beats a dashboard that only reports it.
Read if. Skip if.
READ IF YOU ARE…
- A solo site owner or small-business marketer already paying for Claude who wants ranking, backlink, and technical data without a separate $130-550/month SEO subscription.
- Someone comfortable adding an MCP connector and pasting saved prompt files into Claude, rather than clicking through a dashboard UI.
- A content or SEO person who wants three prioritized action items per report instead of a raw data export to interpret themselves.
SKIP IF…
- You manage SEO for many client sites and need the account-level dashboards, white-label reporting, or team seats Semrush/Ahrefs are built for at that scale.
- You're not willing to pay per-API-call for DataForSEO credits on top of a Claude subscription — this replaces the SEO tool's cost, not all cost.
The full version, fast.
Adobe paid $1.9 billion in cash for Semrush, but the presenter argues most of what Semrush and Ahrefs actually do for a working site owner can be rebuilt by wiring DataForSEO's API into Claude Opus 5 through a custom MCP connector. Once connected, five saved prompt files reproduce site explorer, keyword research, content gap, technical audit, and AI-search-visibility reports — the same functions SEO tools charge $130-550/month for, with AI visibility alone billed as a $60-249/month add-on. The difference isn't only price: because Opus 5 has business context, each report ends with three prioritized fixes and can write or ship the content itself, instead of leaving a table for a human to interpret.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 01:39
01 · Adobe paid $1.9B for this
Opens on the $1.9B Semrush acquisition, walks Semrush's pricing tiers, and previews the four-function rebuild plan (Site Explorer, Keyword Research, Content Gap, Technical Audit) plus a bonus AI-visibility report, framed as 'reporting vs doing.'
01:40 – 04:05
02 · Connecting DataForSEO to Claude
Signs up for a free DataForSEO account, copies the MCP server URL, adds it as a custom connector in Claude, approves the OAuth permission request, and sets tool permissions to Always Allow.
04:06 – 06:41
03 · Rebuilding Site Explorer
Drops the site-explorer prompt file into a Claude project scoped to arankingbot.com; the report returns ranking-keyword counts, average position, estimated traffic value, and three specific pages to fix first with exact edits.
06:42 – 09:02
04 · Keyword research
Runs the keyword-research prompt against a seed keyword, returning 311 collected keywords clustered by intent (informational, commercial, transactional) with volume, difficulty, and AI-Overview-share data instead of a flat spreadsheet.
09:03 – 11:24
05 · Content gap
Compares the site against named competitors to find keyword clusters they rank for and it doesn't, surfacing a specific recommendation to build a free Google-rank-checker tool instead of another article.
11:25 – 12:39
06 · Technical audit
Runs the technical-audit prompt and gets six ranked, real problems (malformed HTML, an oversized /llms.txt file, a stray iframe) instead of hundreds of generic crawler warnings, plus a self-correction to an earlier report.
12:40 – 14:23
07 · AI visibility (the $249/mo one)
Compares Semrush and Ahrefs' AI-visibility add-on pricing ($60 and $249/month respectively), then runs the AI-visibility prompt to show how often the domain is cited across ChatGPT, Google AI Overviews, and other engines, and what to publish to earn more citations.
Lines worth screenshotting.
- Adobe paid $1.9 billion in cash to acquire Semrush, and most of what that tool does can be rebuilt with a DataForSEO API budget plus an existing Claude subscription.
- Ahrefs and Semrush report what's happening on a site; an LLM wired to the same data can also decide what to fix first and write the fix.
- AI search visibility reporting costs an extra $60/month on Semrush's starter plan and at least $249/month on Ahrefs's Standard plan.
- DataForSEO gives $1 of free API credit on signup, but a realistic working budget is a $50 minimum top-up that can last months for a single site.
- Connecting a paid API to Claude takes three steps: add a custom connector, paste the MCP server URL, and approve the OAuth permissions request.
- Setting connector tool permissions to 'Always allow' instead of the default 'needs approval' removes the interruption of confirming every data pull mid-session.
- A five-file prompt library (site explorer, keyword research, content gap, technical audit, AI visibility) turns the workflow into 'paste this file, name the domain' instead of hand-writing a prompt each time.
- A site-explorer report that lists 51 ranking keywords at an average position of 42.7 is only useful once it also names the three highest-upside pages and the exact fix for each.
- Keyword research clustered by intent (informational, commercial, transactional) replaces a flat spreadsheet of keywords a human would otherwise have to sort manually.
- Content-gap analysis can recommend a completely different format — a free interactive tool instead of another article — because the model reasons about the cluster, not just the keyword list.
- A technical audit built this way returned six real, ranked problems instead of the hundreds of generic warnings a crawler-based tool produces, including an oversized /llms.txt file and missing alt text.
- The AI-visibility report tracks literal citation counts across ChatGPT, Google AI Overviews, and Perplexity, showing zero ChatGPT citations against real Google AI Overview presence — a gap invisible to standard SEO tools.
Reasoning over the data beats just reporting it.
WHAT TO LEARN
The same underlying SEO data becomes far more useful once it passes through a model that understands the business, because it turns a raw report into three prioritized, specific actions.
01Adobe paid $1.9B for this
- A single vendor acquisition price ($1.9B) is a useful gut-check for whether a tool's monthly fee reflects data cost or brand/dashboard markup.
- Before replacing a paid tool, separate what it reports from what it lets you decide — the reporting half is usually the commodity part.
02Connecting DataForSEO to Claude
- Custom MCP connectors are added the same way as any first-party integration: name it, paste the server URL, approve the OAuth prompt.
- Setting tool permissions to 'Always allow' instead of 'needs approval' is what turns a connected API from a demo into something usable mid-session.
03Rebuilding Site Explorer
- A ranking report is only useful once it stops at a number and starts naming the specific page and edit that moves it — average position alone isn't actionable.
- Scoping an AI agent's work to a project (rather than a bare chat) keeps a multi-report workflow on one site from bleeding into another.
04Keyword research
- Clustering keywords by intent (informational vs commercial vs transactional) before writing anything prevents wasted content aimed at the wrong stage of the funnel.
- Three hundred related keywords are only valuable once reduced to 'which cluster do I write first' — the reduction is the actual deliverable.
05Content gap
- A content gap isn't always another article — sometimes the highest-leverage gap is a free interactive tool a competitor doesn't have.
- Naming specific competitors up front, instead of letting the tool guess, is what makes a content-gap report point at real, winnable topics.
06Technical audit
- Six ranked, real problems are more useful than a few hundred generic crawler warnings — prioritization is the product, not the warning count.
- A tool that can self-correct an earlier finding (flagging its own prior report as wrong) is more trustworthy than one that can't.
07AI visibility (the $249/mo one)
- Citation count across ChatGPT, Google AI Overviews, and Perplexity is now a trackable metric, and most SEO tools charge $60-249/month extra to show it.
- Publishing original data is what earns citations from AI engines that can't be gamed with backlinks alone — the same logic as classic PR-driven link building.
Terms worth knowing.
- MCP (Model Context Protocol)
- An open protocol that lets an AI model like Claude connect to external tools and data sources — in this video, DataForSEO's live SEO data — through a single server URL instead of a custom integration.
- DataForSEO
- A pay-as-you-go SEO data API that provides keyword volume, rankings, backlink, and SERP data, priced per API call rather than a flat monthly seat.
- AI Overview
- Google's AI-generated summary shown above traditional search results, which can cite a website as a source without generating a click.
- Generative Engine Optimization (GEO)
- The practice of optimizing content to be cited or recommended by AI systems like ChatGPT and Google AI Overviews, as opposed to traditional keyword-ranking SEO.
- llms.txt
- A proposed text file, similar to robots.txt, that some sites publish at their root to give AI crawlers a curated summary of the site's content.
- Content gap analysis
- An SEO technique that compares a site's ranking keywords against competitors to find topics competitors rank for that the site does not yet cover.
Things they pointed at.
00:00toolSemrush
13:13toolAhrefs
Lines you could clip.
00:00
“$1,900,000,000. That's how much Adobe paid to acquire SEMrush in cash.”
00:35
“You're gonna be paying $200 a month for just for SEMrush.”
13:13
“It gets even worse with Ahrefs... I gotta pay at least $249, $250 a month.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogy
$1,900,000,000. That's how much Adobe paid to acquire SEMRush in cash.
In this video, I'm gonna show you how you can replicate all of the functionalities that you would use from an SEO app like SEMRush and even Ahrefs for the cost of your cloud subscription using Opus five. And in my opinion, we're not only going to match the functionalities that this SEO applications give you, but we're going to make it even better.
And I know this might sound a little bit like boy math, but which whichever way you do the maths here, it just doesn't make sense paying for a tool like SEMrush and Ahrefs if you're already using Claude and Opus five. Because if you want the starter pack for doing SEO and AI search, which you need these days, you're gonna be paying $200 a month for just for SEMrush.
Not also the best AI tool out there already. What we're gonna do is replicate the most used functionalities from these applications like SEMrush and Ahrefs. We have the site explorer that tells you, you know, how you're ranking, how many keywords you're ranking for, the average position, volume, few other things, general keyword research, content gap, and a technical audit as well, and even some AI search visibility, which you need these days.
What makes this extra special is that Opus five cannot only do keyword research, you know, backlink analysis, technical audit, everything you need from an SEO tool, but understand something deeper. And it makes it so much better because it's gonna understand your business.
It's gonna tell you exactly what you need to do to fix it because it can reason. It's gonna be able to write content and and even do things for you, like run your Google business profile for you if you wanted to. Anyway, enough of that.
Let's get to actually how we make this happen. For this whole thing to work really well, we need Opus five, and we need a free account with Data for SEO. Data for SEO, um, isn't an application that just give you high quality data, which we can then feed into your AI tool of choice.
This not only works with Opus five, although Opus five is phenomenal for this. You can essentially do this with ChatGPT as well. If you want me to show you how to do that, make sure you leave it in the comments below.
You're gonna create a free account with Data for SEO, and you're gonna have a $1 free credit when you sign up free. I know $1 doesn't seem like it goes a long way, but trust me, it goes a long way, particularly if you're just using it with one business. Um, you have a minimum requirement of $50 credit that you can add from Data for SEO, and I know that sounds a little bit expensive, but that will last you months, if not probably half a year if you're not working on too many websites.
And the good thing here is that it's a pay as you go model. So we need to connect a Dart4SEO to our Claude application.
So Claude has access to real live SEO data, and it's the simplest thing in the world to connect it. What you wanna do is you want to make sure you can copy this, uh, MCP link here.
Right? I'll leave it linked in the video description below so you can access it nice and easily. You wanna go to your desktop application.
You can do this on the just on the browser as well. You wanna go to customize. We wanna go to connectors.
We wanna add a connector here. We're going to add a custom connector. The name, we're gonna name it data for SEO MCP.
And in the remote MCP server, you place in that URL that I just gave you. We're gonna add it in here, and it's gonna ask you to verify the connection. Perfect.
We're going to connect it here. That should pop up your desktop browser that you have already.
And, ideally, you've already you're using the same browser that you logged in to your data for SEO account. It makes it nice and easy. We're going to approve, and that is it.
Right? Now your Claude has access to the highest quality SEO data out there. Now there's one more thing that I recommend you do.
Um, for this, uh, tool permissions, I just say always allow. That shouldn't be by default for you. It'll be on needs approval.
But But I like to say always allow. That way I don't have to get interrupted when I'm doing SEO work with Claude. Now already, because you have OPUS five, which is arguably, you know, the most powerful model out there connected to the SEO data that you need, it's already the best tool out there.
But let's go a step further here. Let's replicate the first functionality, which is the site explorer. So this gives you an overall understanding of where your website is at from an SEO perspective.
Every keyword you're ranking for, the position of those keywords in your website, your average monthly estimated search volume, plus a few of the pages you should probably start fixing right away. In the in the video description below, you'll get a link to access all of these prompts that you can use to follow along.
They're gonna be available in our free community. If you're on a mobile and you don't see the MD files, please swipe your phone to the right. I can't believe how many people don't think of that, and they tell me that the files aren't there, and I've somehow cheated them.
The files are there. You can download for free. I'm not asking for any payment.
Anyway, once you get this, we can use the first file here, the site explorer. You wanna copy that, um, MD file here, and then you wanna make sure that ideally you've started a project named the website that you're working on.
And I'm gonna say, perfect. Run this on the website and then put the website URL.
That's all you need. Right? You know, run that.
They're gonna give it a couple of seconds. Now Claude Opus, make sure you've got that selected here. It's gonna do a lot of things here.
It's gonna search the website. It's gonna understand the context, and then it's gonna do keyword research. This is why I think this is by far a lot better than a Ahrefs or SEMRush.
And after a couple of minutes, because we have used the artifacts feature to build our kinda dashboard, which we can reuse if we want, we've got here the site explorer report. It gives me tells me that I'm ranking for 51 keywords. I've got 42 monthly estimated search volume, and it gives me a bit of a strategy here.
The domain appears twice in the rankings, the top pages by the estimated organic traffic.
This is a blog post in the SEO audit, which is right. Keywords doing the work and a few more things here. Three pages with the most realistic ninety day upside.
Great. So it's already telling me, hey. These are the things you should probably start working on.
Um, let's see what else where the competition is sitting at. And it tells me exactly the first three things that I need to do. Fix the home page h one, rewrite this page, and it even tells me what to do, and update the personalities post, um, to make it more update.
Perfect. Number one, already done. Right?
Now you can even connect Claude or Opus five to your WordPress account via the MCP or depending on how you build your website. There's a very large probability that you can connect Opus five directly to your CMS so you can ask it to do the work for you. But let's just stick with SEO components for now.
The keyword research. Right? A lot people love to overcomplicate this, and keyword research is still somewhat important these days, particularly to find out what are the transactional keywords that people are looking for.
This then also kinda gives a way to AI search and getting cited and getting recommended by the AI search engines. But let's go to number two. And, again, in the resources, there's an MD file keyword research.
You just gotta do the same thing. We're gonna drop the file in there and say, run this for the site that we're doing work for. And, again, after a couple of minutes, we got the keyword research, what to build in what order.
I gave it the seed keyword. If your seed keyword is different, obviously, tell it the one that you wanna concentrate on. It'll ask you a couple of questions like, you know, which keyword?
Are you sure if it's this one? Give you a couple of recommendations and also, um, kinda clarify which market.
If it's The United States or in Australia, Canada, or if you're speaking another language, it'll kinda tell you that as well. And then it's pulled a bunch of the keywords that it recommends one finding that you should change what you write.
Okay. Cool. It's telling me the keywords I should probably concentrate on group one informational.
So okay. I need probably start writing a lot more helpful content about this.
So what is generative engine optimization, GEO, for example? What is AI SEO? SEO for GPT, SEO for AI search and AI overviews.
These these are really good. And then I've got the commercial investigations, so the more transactional based keyword.
These are the ones that are gonna actually give me money. And I didn't really have to, you know, filter and kind of dig and waste hours trying to find which keyword is gonna be best for my business and which one's transactional, which one's informational, and how do I kinda class them all together because it has the con business context, then it kind of, for the most part, knows exactly what it needs to do.
Don't trust it blindly, but already I've got a pretty solid keyword research document telling me exactly what I need to do and the first three things to do as well. If you it also tells you where exactly got the sources, obviously, from data for SEO because we've connected it made that connection previously.
You can kinda rest assured that the data it's getting is high quality. And when you have high quality data, an incredible smart model like Opus four or an incredibly smart model like Opus five, then you're kind of .
Number three, the content gap. So everything that the competitors are ranking for and you are not. Now you might not always want this, but it's a great way to start.
It's also a good way to understand where how much maybe further, uh, away your competitors are in terms of competition, but it's also a great way to find keywords that you need to rank for that you probably didn't even didn't think of.
So let's get to the third one. Same thing here. Again, within the prompts that will be in the free community, there's the content gap.
We're gonna drop that little guy in there. If you have an understanding of exactly who your competitors are, then say it here in this section, uh, because otherwise, it's going to see which other website is overlapping in the amount of keywords, um, the most. So if you do have an understanding who that is, tell it in here.
I want you to do the content gap up with these competitors. That's always really good option. I'm gonna hit enter here.
And after a couple of minutes, here's the content gap. Finding that should decide the next quarter. This one's really interesting.
It's telling me to that I need to generate free tools, um, which is easy enough for me to do, but I didn't really think about this before. And I think this is the power of Opus five, um, having the right context.
So the context from the high quality data, but also about what the business is about. Right? So tell me, uh, do a free tool, not an article.
So my Google ranking, how can I check my Google ranking? How to find your Google ranking? So it's telling me, cool.
Nothing clever. It needs to be a tool. This cluster is the single clearest, uh, format for the whole dataset.
So, essentially, I need to build a tool that will allow us to the user to put in their website URL, and I show them how they're ranking or what their Google ranking is.
Easy enough to do. I love creating tools like that. That's probably gonna give me not only a lot of traffic, but also a lot of backlinks as well.
And this is another cluster that I need to write about because my competitors are ranking for one. I'm not the AI SEO software. That's a good one.
A cluster dedicated to other free tools as well. This is a good one as well. Classified SEO's surfer SEO alternatives.
Small, but pure buying intent. That's not a bad idea. Creating a good cluster of, you know, all the other SEO tools that we're that we're comparing to.
I like this already. Do you see how like, it's just so much smarter than all the other SEO tools, I think, and and you're paying peanuts for those calls from data for SEO.
The technical audit is one that I love to do because it tells you, you know, how your website is performing. And a lot of people don't realize that it might be too slow. You don't have the right schema.
You don't have the right keywords selected in here. This will kinda tell you exactly what you need to do. So, again, you drop that prompt in there.
You tell it which side you wanted to work on. You hit the go button, and it gives me some good understanding back.
So my on-site SEO is okay. Right?
So winning the site is unusually good technical shape as it should be, but it's got so a lighthouse report of a 100. It's loading very, very quickly. And I got a few corrections here.
Let's see. Some home page images are missing the alt text. That's something I should fix.
Particularly, I go on about that a lot, so I shouldn't have kinda mixed missed that. But, obviously, you know, it's not gonna be a 100% perfect. The reload text HTML ratio, didn't really even think of that, but that's a good point.
Fixed eventually. The LLM dot TXT file, iframe, things to look at as well.
So nothing massive, but here's the first three things that I recommend I need to fix from an on-site SEO perspective. That test alone is going to be vastly better than any on-site SEO tool that you could get from a Ahrefs and SEMrush. And last one, and by far, the kind of more interesting and money saving one is the AI visibility test.
Right? This is incredible because if we just look at the pricing of what Ahrefs and SEMrush expect you to pay just to get access to AI visibility.
So it's in the starter pack for SEMrush. And, you know, if I go from the normal SEO pack to the starter that has the AI search, I would have to pay an extra $60, essentially.
It gets even worse with Ahrefs. This is the pricing page because if I want access to any AI suggestion stuff, I gotta pay at least $249, $250 a month.
But with this connection, you put in the AI visibility MD file. You hit enter. It gives you a complete AI visibility report telling you where the AI is citing you.
So no GPT citations, a couple of Google AI citations, which is good. So I'm actually am getting cited on GPT and Google AI overviews. Gives you the competitors, the scoreboard, the one page that is working really well, and therefore, how we need to kind of replicate that good performance, and the three things I do first to rank higher on GBT and Google AI overviews.
That functionality alone, Ahrefs, SEMRush are expecting you to pay at least another $60 for. Now you have access to it for the same amount that you're already paying for your Claude subscription. If you're watching this video, I'm assuming you're already paying for a subscription for Claude anyway.
Again, all the resources are gonna be available in the free AI ranking school community. If you're on mobile, if you don't see the downloads, for the love of god, please swipe to your right. I hope you enjoyed that.
I hope you save a few bucks a month because you don't don't have to pay for Ahrefs and SEMrush anymore. You just gotta play for your anthropic subscription.
Again, if you want me to do this video with GPT 5.6, let me know in the comments below. Thanks.
The Hook
The bait, then the rug-pull.
Adobe just paid $1.9 billion in cash for Semrush — and this video spends fourteen minutes rebuilding the parts of it a working site owner actually uses, for the price of API credits and a Claude subscription already being paid for.
Named ideas worth stealing.
00:48list
Four Functions, Live
- Site Explorer
- Keyword Research
- Content Gap
- Technical Audit
The four SEO functions the video sets out to rebuild live using DataForSEO + Claude, plus a bonus fifth: AI search visibility.
Steal forany tool-replacement pitch — name the exact features being displaced before showing the build
01:20model
Reporting vs Doing
A capability comparison table splitting SEO tool functions into two categories — telling you what's happening (keyword data, backlinks, technical crawl) vs doing something about it (deciding what to fix, writing the content, fixing internal links) — with Semrush winning every 'telling' row and Opus 5 winning every 'doing' row.
Steal forpositioning any AI tool against an existing category leader without competing on raw data coverage
How they asked for the click.
VERBAL ASK
14:23link
“All the resources are gonna be available in the free AI ranking school community.”
Soft CTA delivered in the close, paired with an on-screen screenshot of the Skool community; ends with a comment-bait ask for a follow-up video built on a different model.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
AFFILIATECommission earned if you click.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
