Turn a spreadsheet of URLs into tracked, tagged pages.
Hand your Page Pulse MCP a CSV — competitor pages, vendor profiles on a review site, a whole client portfolio — and it creates one project per domain, tracks the exact URLs (no messy discovery crawl), and tags each one by type and category, in a single guided pass.
A hundred URLs is a hundred trips through the same four clicks
Standing up monitoring for a list of pages by hand is the same tiny task on repeat: create the project, add the page, add the tags, move on — times however many rows are in your sheet. It is slow, and it is where consistency quietly dies. Half your competitor pages end up tagged Pricing and half pricing page, and three months later the tag-based alert you set up only catches half of them.
This workflow does the whole list in one pass. It reads the sheet, tells you up front how many projects and pages it will create and where the data is messy, and only builds once you approve. It tracks the exact URLs you listed — not a discovery crawl that floods a competitor’s domain with hundreds of irrelevant pages — and it applies your labels consistently, which is the whole point.
The tags are the payoff. Once every page is labeled, alerts and reports scope by tag name across every project at once — so you watch every “Pricing Page,” or every competitor in one vertical, without touching a single project setting again.
The prompt
Connect the Page Pulse MCP to Claude or ChatGPT, paste this in, and it builds the report from your live data.
Swap the two label columns for any pairing — page type × product line, content type × region, competitor × vertical. Tags are a flat list per page, so extra columns just mean extra tags; tell it which columns to use.
Attached is a CSV of URLs I want to track in Page Pulse. Before you create anything, confirm you can do this correctly: walk me through your plan and flag any data problems first. Don't create anything until I say go. THE DATA - Column [B] holds the URL. - Column [C] holds the [PAGE TYPE]. - Column [D] holds the [CATEGORY / VERTICAL]. - Ignore [header rows, example/demo rows, anything above row N]. WHAT I WANT - One Page Pulse project per website. Pages must live on the project's own domain — different domains cannot share a project. - Track only the exact URLs in the sheet: use URL-list mode, NOT a site-discovery crawl. - Tag every URL with BOTH its [PAGE TYPE] and its [CATEGORY]. BEFORE YOU RUN, TELL ME 1. How many distinct domains (= projects) and how many URLs total. 2. Any duplicate URLs, especially rows with conflicting labels. 3. Any inconsistent or misspelled labels, and how you'd normalize them. 4. Any project names you can't confidently derive from the domain. HOW TO RUN IT - Names must be right the first time (there is no rename tool), so confirm any business name you can't read off the domain before creating. - Create in URL-list mode, then tag each page-type group in its own add-pages call — every URL in one call shares that call's tags. - Do a test batch of 5 projects first, verify the tags landed, then stop and check in with me before doing the rest. - Report progress as you go. If you hit a limit, tell me exactly where you stopped and what's still pending. I should be connected to the [ORGANIZATION] org — confirm what you see before creating anything.
How to run it
One guided pass, with a checkpoint before it touches your org.
Get your sheet ready
One row per URL, with a column for the URL and two columns for your labels — e.g. page type and category/vertical. It does not need to be tidy; the prompt asks the model to flag duplicates and inconsistent labels before anything is created.
Connect the MCP and paste the prompt
With the Page Pulse MCP connected, fill in the [bracketed] parts, attach the CSV, and send. It confirms which organization it sees, then reports how many projects and URLs it will create and where the data is messy — before building anything.
Approve, then run a test batch of 5
Give it the go-ahead. It creates the first 5 projects in URL-list mode and tags each page group, then stops so you can confirm in the app that the pages and tags landed the way you expect.
Let it finish the list
Once the test batch checks out, it works through the rest, reporting progress. Budget roughly two tool calls per URL group (one to create, one to tag), so a long list will span several turns — it will tell you exactly where it stops if it hits a limit.
What it creates
- One project per domain (pages can't span domains)
- Exactly the URLs you listed — no discovery crawl
- Each URL tagged by page type AND category
- Re-listing a URL just stacks its tags, never errors
Know before you start
- New projects are crawl-only — no traffic data
- GSC/GA4 needs a Google sign-in in the web app (not via MCP)
- No rename tool — project names must be right first time
- First crawl runs on the nightly cycle; changes show from the 2nd
Why the tags matter
- Alerts and reports scope by tag name across ALL projects
- Watch every "Pricing Page" in one alert
- Or every competitor in one vertical at once
- No per-project configuration — tag once, scope forever
Track your whole watchlist in one pass
Connect the Page Pulse MCP, hand it your list, and let it build the tracked, tagged projects for you. Free for up to 50 pages.
No credit card required · Free plan available
Related workflows
Compare GSC & GA4 performance before and after any page change
Stop guessing whether that rewrite worked. Record automatic, daily, page-level changes, then pull a clean before/after comparison of…
View workflow
Find which landing pages bring in the most key events from GA4
GA4 shows you key events, but not which landing page caused them. This workflow surfaces the exact pages…
View workflow
Find which landing pages drive your most valuable conversions
Not all landing pages are equal. Some drive clicks that convert while others drive clicks that bounce. This…
View workflow