Connect Search Console and MetaPilot ranks the pages Google is already showing people that still have no structured data, highest impressions first. Optimize one and the queries it ranks for go into the prompt that writes its title, instead of a keyword guessed from the copy.
Ranked by the traffic they already earn. Optimizing the first one writes its title from the query underneath it.
TL;DR
Google already knows which of your pages it puts in front of people, and which of those say nothing a machine can read. Connect Search Console, read-only, and MetaPilot ranks exactly those pages by impressions, then puts the queries a page already ranks for into the prompt that writes its title, instead of a keyword inferred from the copy. It is on every plan including Free, and it is not a rank tracker.
You have more pages than attention, so something has to decide which one gets optimized next. Usually that something is a hunch.
Google answered the question a while ago. It is putting certain pages of yours in front of real people right now, and some of those pages say nothing a machine can read: the guide, the collection, the landing page your platform never covered because it only understands products.
The same data settles a second question. Writing a title means choosing the phrase to lead with, and read off the page that choice is a guess. The phrase people actually typed has been sitting in Search Console the whole time.
Settings → Search Console, one read-only grant, then pick one of your verified properties. A domain property is the recommended choice: it covers http, https, www and non-www together, so more of your pages match.
Every page Google sends impressions to that still has no structured data, ranked by impressions, each row showing the query it already ranks for.
The page goes through the same pipeline as everything else, review step included. What is different is the prompt: it carries that page’s own queries and click-through rate.
Writing a title means picking the phrase to lead with. Read off the page, that is an inference. Google has the measurement, so once you are connected the searches a page already ranks for go into the generation prompt and the highest-impression one is the phrase the title leads with.
Not by crawl order, not by a hunch. Impressions first, so the top of the list is the page costing you the most by staying invisible to machines.
Every other reason to rewrite is structural: missing, truncated, duplicated, keyword-stuffed. This one is a number. High impressions with a click-through rate well below what that position would ordinarily earn is a real problem, and it is invisible from the page itself.
A closed platform covers the page types it understands. The guide, the collection, the landing page are the ones that show up here earning impressions with nothing machine-readable on them, and structured data is one click away on each.
The search signals are built per generation, so any page MetaPilot writes for you while you are connected carries its own numbers into the prompt. That includes auto-pilot regenerating a page after your content changed.
MetaPilot never writes to your Search Console, submits a sitemap, or requests indexing. Disconnecting withdraws the access at Google and keeps your history here, so reconnecting later finds the data waiting.
Search Console is on every plan, including Free. There is no connect gate and nothing extra to buy: a Free organization that can see which of its pages already earn impressions has a concrete reason to grow, and a locked door is just an objection.
One thing scales with your plan: how much history is kept. Search rows are held for 7 days on Free up to 365 days on Business, the same window that covers your AI-crawler analytics and uptime samples. Your first sync backfills as much history as your plan keeps.
No. This one is a read from Google over OAuth rather than something happening in your site’s request path, so connecting and working the opportunities list is identical whatever your site runs on. What still depends on your platform is delivering the finished metadata to your pages: the capability matrix is the honest answer there, and it is generated from the code.
No, and deliberately not. Google already gives you impressions, clicks and position for free, and a worse copy of Search Console is not worth your time. What this is for is deciding what to optimize next, and making the writing better when you do. Top pages and top queries are in the product as reference, sitting below the opportunities and collapsed.
Google finalizes a day two to three days after it happens, then keeps revising it, so the most recent days are always provisional. Every run re-pulls the last 5 days and replaces those rows, which is how the numbers correct themselves. Nobody can give you real-time Search Console data.
Google withholds anonymized queries, the rare ones that could identify the person searching, so the queries listed for a page always sum to less than the page’s own row. That is Google’s behaviour rather than a sync problem, and it is true in Search Console itself.
Read-only Search Console access, and nothing else. We list your verified properties so you can pick one, and we read search analytics for that property once a night. You choose the property yourself: a Google account often holds several, and picking the wrong one silently reports on the wrong site.
Connect Search Console, sort by impressions, and work down the list. It is free on every plan, and read-only.
Ask an AI whether search data belongs in your metadata