Overview

Search data: where it comes from

Which sites does it cover?#

Docsbook reads a Search Console domain property for docsbook.io, which covers every subdomain, and keeps the rows whose URLs belong to your site.

Where your site is served Search Console data
<owner>.docsbook.io/<repo>, built from your GitHub repository Read automatically
Its own <name>.docsbook.io address, hosted by Docsbook or set in Settings ▸ General ▸ Site source Not read yet
Your custom domain, such as docs.example.com Outside Docsbook's property

For a custom domain, verify the domain in your own Search Console account to see its numbers there.

What syncs, and how often?#

Docsbook asks Google for two views of the same traffic and keeps both.

  • Queries — each query Google names for your pages, per day and per page, with clicks, impressions and average position
  • Page totals — the same numbers per page, including traffic behind queries Google withholds for privacy
  • 30 days — how much history is kept
  • About two days behind — Google's own lag; a "data through" date shows how fresh the numbers are
  • Once a day at most — a pull happens when the agent needs rankings, or when you press Refresh on the Search rankings card in the panel chat

Access is read-only. Docsbook never submits URLs, requests indexing or changes anything in Search Console.

How to find the pages closest to page one#

1
Open Analytics ▸ SEO

The list opens on the Queries view: each row shows clicks, views and CTR, with the average position as a chip such as #6.4.

2
Switch the View to Pages

Each page shows the same numbers and how many queries reach it.

3
Sort by position, best first

Read down to the rows between #5 and #20 with many views. Google already shows these pages; a better title or first screen is what wins the click.

What does the agent do with it?#

The Docsbook agent uses the same rows to choose search changes and to judge them:

  • Picks the pages worth fixing — queries at positions 5–20, already visible and not yet winning the click
  • Forecasts in clicks — for a title, snippet, intent or first-screen change, it compares the page's CTR with what your other pages earn at that position and forecasts 25–50% of the gap, labelled an estimate
  • Learns from your site — once five of your own changes have been measured, that band is replaced by what they actually achieved
  • Refuses to guess — under 30 impressions in 28 days a page gets no forecast, never a zero
  • Judges the change — Google's position, impressions and clicks for the page, before against after, with the day of the change left out
  • Explains drops — the Explain a traffic drop trigger names the pages and sources that lost readers

What about Bing and other engines?#

Search Console only reports queries you already appear for, so Docsbook also checks the results pages themselves. Every day it reads Google and Bing for up to 5 queries per engine, including queries you don't rank for yet; choose them by asking the agent or with configure_mentions. The results land in Analytics ▸ SEO ▸ Google & Bing mentions.

You can also push your own daily numbers from Bing, Yandex, DuckDuckGo or a rank tracker, using the project API key from Settings ▸ Domain & API:

curl -X POST https://docsbook.io/api/v1/search-engine-stats \
  -H "Authorization: Bearer <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{"engine": "bing", "date": "2026-09-20", "query": "docs hosting", "page": "/quickstart", "views": 120, "clicks": 9, "position": 6.2}'

The rows appear beside Google's in Queries and Pages, marked with their engine. Send one row per call, or up to 200 as {"stats": [...]}.

Field Value
engine google, bing, yandex or duckduckgo
date The day the numbers belong to, as YYYY-MM-DD
query The search query
page The page it led to (optional)
views, clicks, position The engine's figures (optional)

FAQ#

Do I need to verify my site in Google Search Console?

No, for a site served at <owner>.docsbook.io/<repo>: Docsbook's own property already covers it. A custom domain needs verifying in your own account.

Why are these numbers lower than Search Console's own totals?

Google leaves rare queries out of its query report for privacy, and the Queries and Pages views are built from the queries Google names. The agent's forecasts and before-and-after checks use the page-level totals, which include that traffic.

Why don't I see yesterday's searches?

Google publishes Search Console data about two days late. The newest day shown is the one Google has closed.

Next steps#

Updated

Was this page helpful?