Search data

What is search data in Serpel?

Rank tracking checks the keywords you chose. Search data shows every query that Search Console reports for your site (Google leaves out rare ones), with the clicks you received. Together they answer two questions: where do I stand for the keywords I care about, and which searches already bring visitors that I don’t track yet? The search data feature page gives the overview, this page explains how it works.

How do you connect Google Search Console and Bing Webmaster Tools?

You can connect once for your whole account or separately for one project. The account route is faster, because Settings → Integrations matches the projects that you own by domain. The project route is the Connect button on the Search data page, and it is the only route for a project where you are an editor. Owners and editors can connect, viewers can’t. The website must already be verified with the provider.

1. 1

   Open Settings → Integrations.

   The first section, Search data and web analytics, has a card for Google and one for Bing Webmaster Tools.

2. 2

   Press Connect.

   For Google, approve the read-only request in the browser. For Bing, paste the API key and press Connect. Serpel stores credentials encrypted.

3. 3

   Check the assignment.

   The account card shows “Assigned to N projects”. Open Search data in a project to see the property or website, the source of the data and when it was last fetched. If one project has no clear match, the page asks you to choose: press Manage and pick the Property (Google) or Website (Bing).

The Google sign-in on the account card asks for Search Console and Google Analytics together, so the same connection also serves web analytics. If you grant only one of the two, the card shows “Only Search Console granted” or “Only Analytics granted”.

How are connections assigned to projects?

When you connect an account integration, and whenever you create a project, Serpel looks at the domain of every project that you own and picks the matching property or website. Projects where you are only an editor aren’t matched, so connect them in the project. Match again in the integration dialog repeats this after you verify a new site. A connection you made by hand in a project stays untouched and is listed as “connected manually in the project”.

• Disconnecting an account integration deletes the stored credentials, and every project connected through it loses this data until you reconnect.
• A project connection that comes from the account can only be removed on the account card. A manual project connection has its own Disconnect button under Manage.
• If Google revokes access, Reconnect asks for approval again, on the account card or under Manage for a manual connection.

What does the Search data page show?

The toolbar holds the Google and Bing switch, the period (Last 7 days, Last 28 days, which is the default, and Last 3 months, which is 90 days) and, for Google, the search type: Web, Images, Videos, News and Discover. Four figures lead the page, Clicks, Impressions, CTR and Avg. position, each compared with the previous period of the same length. The providers deliver data with a delay of two to three days.

Below the figures, cards group the data: Queries, Pages, Countries with a globe, Devices with Search appearance, Query types and Position distribution. Click a country, device, search appearance or folder to filter the whole page, and remove the filter from the bar above. Click a query or a page to open its details: totals, the pages or queries behind it, countries, devices and your tracked keywords. The All search queries list shows the top 100 queries by clicks with views for page 1, page 2 and queries you don’t track yet. The Track button on a row adds the query to rank tracking. Discover has no queries, because Google shows those pages in a feed without a search.

Which insights does Serpel add?

Serpel analyses the web search results of the period for you, and filters don’t narrow the insights. Each insight looks at the top 5,000 queries by impressions, or the top 5,000 pages for Low CTR on the Pages card. If Serpel has to read the data live from Google, because its stored history doesn’t cover the period yet, the limit is 500 rows. Opportunities and Low CTR also need enough impressions: 38 in 7 days, 150 in 28 days and 482 in 90 days. Position distribution and Query types count every analysed row, however few impressions it has.

Opportunities combines the Search Console data of the last 28 days with your rankings, crawl and codebase to rank what to fix first. The insights on the Search data page are the quick look, the opportunities page is the to-do list.

How do the index and crawl views work?

Switch the toolbar from Performance to Index and crawl to see how the provider treats your pages.

• Google: Serpel inspects up to 20 of your top pages by clicks of the last 28 days with Google’s URL inspection. The URL inspection card has the tabs All, Not indexed and Errors, and a Sitemaps card lists your submitted sitemaps. Google allows 2,000 inspections a day and 600 a minute per property. Results are kept for 24 hours. Pages beyond the quota are skipped and the page says so.
• Bing: a chart Bing crawl and index with crawled pages, indexed pages, crawl errors and inbound links, a Crawl issues card with 9 issue types and their severity (new URLs are flagged), Sitemaps and feeds and Inbound links per page. Bing data appears after the first daily sync, within about an hour of connecting.

How does URL submission to Bing work?

Serpel never sends URLs to Bing on its own. To allow it, open Manage for Bing and switch on Allow Serpel to submit URLs to Bing. The URL submission card then shows Write access on, and Submit URLs opens a dialog where you enter one full URL per line. Every URL must belong to the connected website, you can send up to 500 URLs at a time, and URLs you submitted in the last 24 hours are skipped. Serpel checks Bing’s daily and monthly quota first and submits nothing if it is too low. Only owners and editors can submit.

How often does Serpel sync and how long does it keep search data?

• Google: the first sync imports the last 90 days. After that, a sync runs about every 23 hours and refreshes the last 3 days, because Google still adjusts recent numbers. Serpel stores up to 5,000 query rows per day and keeps 16 months. A period is answered from this history when it covers the whole period, and live from Google otherwise.
• Bing: a daily sync stores the crawl history, crawl issues and inbound links. Clicks and impressions are read from Bing and cached for six hours. Bing delivers its data with a delay and partly only weekly.

Every sync appears as Search data sync or Bing index sync under History (see history and exports). If a sync fails, the connection strip on the Search data page shows an error badge whose tooltip names the last error.

How do you export search data to CSV?

Press Export at the top right of the Search data page. The menu offers All search queries, Pages, Countries and Devices (Bing has only queries and pages). The file is a semicolon-separated CSV with a UTF-8 byte order mark, named like search-queries-2026-11-05.csv. The query file has the columns Search query, Clicks, Change in %, Impressions, CTR in %, Avg. position and Tracked. It contains what the page shows, so up to 100 rows per list. For up to 500 rows, use the CLI or API.

How do you use search data from the CLI, API or MCP?

The CLI and the REST API cover connecting, reading and the Bing tools. Reading needs the scope projects:read, connecting and submitting needs projects:write. --days takes 7, 28 or 90, --filter takes values such as country:DE or device:mobile, and --provider bing switches the source.

serpel search status --project "$PROJECT_ID"
serpel search connect-google --project "$PROJECT_ID"
serpel search connect-bing --project "$PROJECT_ID" --api-key-file bing.key
serpel search queries --project "$PROJECT_ID" --days 28 --limit 50
serpel search breakdown --project "$PROJECT_ID" --dimension country --filter device:mobile
serpel search opportunities --project "$PROJECT_ID"
serpel search index --project "$PROJECT_ID" --limit 20
serpel bing issues --project "$PROJECT_ID" --severity error
serpel bing write-access --project "$PROJECT_ID" --enable --yes
serpel bing submit --project "$PROJECT_ID" --url https://www.example.com/new-page

curl "https://serpel.app/api/v1/projects/$PROJECT_ID/search-breakdown?provider=google&dimension=query&days=28&limit=25" \
  -H "Authorization: Bearer $SERPEL_TOKEN"

curl "https://serpel.app/api/v1/projects/$PROJECT_ID/search-insights?provider=google&kind=opportunities&days=28" \
  -H "Authorization: Bearer $SERPEL_TOKEN"

For AI assistants, the MCP tool get_search_performance returns the totals, the change against the previous period and the top queries or pages, at most 25 rows per call. It reads the same connection and needs no credits. Setup is on the MCP page for Claude Code.

Where to go next