When I announced the second generation of VitalSentinel last month, I ended with two things I was genuinely excited about and had to admit were not shipped yet: a public API, and an MCP server.
Both are live today, and the MCP server is the one I want to start with, because it changes how you use VitalSentinel day to day. MCP is simply the standard way an AI assistant connects to a service. If you already use Claude or ChatGPT, connecting VitalSentinel feels like linking an account, and takes about as long.
Going to Look, Versus Just Asking
All eight modules already report into one place, so your data has never been far away. But looking is still looking. You open the app, pick the domain, pick the range, find the chart, read it. That is exactly what you want when you are studying something properly.
It is more than you want when you simply have a question. Most questions are small – is anything down, did that page get slower, are we still indexed – and the answer is one number you wanted ten seconds ago. The friction was never in the data. It has been sitting there, correct and complete, the whole time. The friction is in going to get it.
Now You Just Ask
Connect VitalSentinel to Claude once, and then ask:
- "Which pages regressed on LCP this month?"
- "Did anything go down last night?"
- "Which queries lost clicks after the redesign?"
- "Is anything blocked in robots.txt that shouldn't be?"
And you get an answer. Not a link to a dashboard, an answer. Three product pages slipped on LCP, here they are, and here is the week it started. Yes, there was a four-minute outage at 02:14, and while we are here, a certificate on that domain expires in nine days. These six queries lost the most clicks, and they all sit on the same template.
It is your real, live data behind every one of those, which means "why?" is now a reasonable follow-up. You can keep pulling on the thread in the same conversation instead of opening another tab and starting over.
It works the same way in Claude, ChatGPT, Claude Code, Cursor, VS Code, and Notion. Whichever assistant you already have open is the one you ask.
Connecting Takes Under a Minute
In Claude or ChatGPT, open your assistant's settings, add a custom connector, and paste in this address: https://mcp.vitalsentinel.com/mcp. Sign in with your browser when it asks, and you are done. Cursor and VS Code take the same address.
If you use Claude Code, it is one line:
claude mcp add --transport http vitalsentinel https://mcp.vitalsentinel.com/mcp
There is nothing to install and nothing to keep updated afterward. No app, no download, and no key to generate and paste anywhere – you sign in with your browser, the way you sign in to anything else. There is no setup screen inside VitalSentinel to go hunting for either, because you connect from the assistant's side.
The connection guide has the steps for each assistant, and there is a dedicated MCP page here if you want the overview first.
What You Can Ask About
There are more than 50 things the assistant can reach in your account. In plain terms, that means your Core Web Vitals and real-user performance, whether your sites are up and your certificates still valid, your scheduled Lighthouse runs, your search traffic and the queries behind it, what is indexed and what is not, your robots.txt and sitemaps, your analytics and ecommerce numbers, your alerts, and your reports.
This is not a new module in the product. It is a new way into the eight you already have.
You never have to know which piece it needs. You ask about LCP on your product pages, and it goes and fetches the vitals and the timeline behind them, which is exactly the pair of screens you would have opened yourself.
There are also five ready-made prompts for the checks you repeat: a weekly performance review, a traffic-drop investigation, a pre-deploy check, a Core Web Vitals triage, and an incident postmortem. Pick one instead of describing the whole routine again.
If You Came Looking for a Search Console MCP
That is the one people ask for most, and it is included. Your clicks and impressions, the queries behind them, the pages they land on, what Google has indexed and what is stuck outside it, your sitemaps and your robots.txt are all reachable, because Google Search Console is already one of the modules. Connect it to VitalSentinel once and every assistant you connect afterwards can read it. There is no second connector and nothing extra to authorize.
What you get on top of a Search Console connector is the context around it. The ranking drop and the LCP regression that might explain it are in the same conversation, so "did that drop line up with the slowdown?" is one question rather than two exports and an afternoon lining them up by hand.
You Stay in Control
I want to be straight about this rather than pretending it is look-but-don't-touch. A connected assistant can act on your behalf, not just read to you, and that is most of what makes it useful. Three things are what keep it safe.
It holds no credentials of its own. No database access, no service account, no shared key sitting somewhere waiting to leak. It passes your own sign-in through and stores nothing.
It can never do more than you can. What a connection is allowed to do is your own role in the workspace, rechecked on every single request. If your access changes, the assistant's changes with it, immediately.
And you decide what it reaches. The approval screen shows you everything being requested, grouped by how sensitive it is, and you choose which workspaces it can reach before you approve. You can disconnect it whenever you want.
There Is an API Too, on Every Plan
If you would rather have the raw data than a conversation, there is a public REST API as well, at https://api.vitalsentinel.com/api/public/v1. You create a key in the app under Settings, then API access, and the API introduction covers the rest.
Neither the API nor the MCP server sits behind a plan tier. Both work on every plan, including Free. Usage runs on credits per billing cycle:
- Free – 500
- Starter – 2,500
- Lite – 25,000
- Standard – 100,000
- Advanced – 250,000
You pay for capacity, not features, same as everything else here.
Where to Start
Connect it and ask it something about a domain you already monitor. That takes about a minute and will tell you more than the rest of this post did.
If you would rather read first, the MCP introduction covers the assistant side. Everything you can already see in RUM Monitoring, Uptime Monitoring, Google Search Console, and Web Analytics is now a question away, on any plan.
Two things I promised in July, shipped. Go connect it, and tell me what you end up asking it.
Written by
Martin Stepanek
Enterprise Technical SEO Consultant
15+ years building enterprise websites and designing the infrastructure behind them, and 5+ in technical SEO and web performance. My work is making companies findable – not just on Google, but inside ChatGPT and the other AI engines people now ask first. I write a regular newsletter and blog on technical SEO at TechSEO Vitals, and I built VitalSentinel because the gap between something breaking and someone noticing should be minutes, not days.
