Data to AI™ / Gemini/Google Search Console
Bring search performance into the codebase where URLs actually change, so site migrations, internal links and SEO snapshots are built from real Clicks and Queries.
Better answers, fewer tokens.
Gemini CLI
Which blog pages lost the most clicks in September compared with August?
The CRM guide lost 1,240 clicks, down 37%, the largest drop, and now averages Position 7.8. The other two pages fell 21% and 12%.
| URL | Clicks (Sep) | Clicks (Aug) | Position (Sep) |
|---|---|---|---|
| /blog/crm-guide Daily Metrics by Page | 2,140 | 3,380 | 7.8 |
| /blog/pricing-models Daily Metrics by Page | 1,620 | 2,050 | 5.2 |
| /blog/api-basics Daily Metrics by Page | 980 | 1,110 | 6.4 |
Page rows are sampled by Google, so pages are compared with each other and not added up into a site total.
Google Search Console shows how a site performs in Google search: Clicks, Impressions, CTR and average Position by date, query, page, country and device, for the properties your account can access.
Query Daily Metrics by Page through the dataddo server, write every URL with Clicks in the last 90 days to migration/urls-with-traffic.csv, then list the ones missing from redirects.yaml.
A web developer runs this before merging the new URL structure. Pages that earn organic traffic get a redirect first, and the CSV goes into the pull request as evidence.
From Daily Metrics by Page and Query, take the top Query by Clicks for each URL under /guides and save the pairs to data/anchor-text.json.
A technical SEO feeds the JSON into the script that suggests internal links, so anchor text follows what searchers type rather than what the team guesses.
Using Daily Metrics, write reports/seo/weekly.md with Clicks, Impressions, CTR and Position per Date for the 7 days ending 4 days ago, and compare with the week before.
A marketing engineer runs the prompt in Gemini CLI every Monday. Each snapshot is committed, so the history of search performance sits next to the site code.
Data to AI™ is priced per seat, like the AI assistants your team already uses, with ~25% off on annual billing. Need tailored contracts or volume pricing? Talk to us about Enterprise.
Serve governed business data to Claude, ChatGPT and other AI tools over MCP.
$10.49/month
per seat
3 seats minimum
Flexibly deploy Dataddo in any cloud or hybrid environment.
Custom
We have a payment model that works for you
| Without Dataddo |
With Dataddo
|
Outcome for you | |
|---|---|---|---|
| Migration safety | Redirect lists are built from a sitemap, so old URLs that still earn search traffic but were never in the sitemap are lost. | Gemini reads every URL with Clicks or Impressions from Daily Metrics by Page and compares it with the redirect file in the repo. | Traffic kept after a migration |
| Processing delay in automated runs | A script that asks for yesterday gets numbers Google is still processing and reports a false drop. | The source reads a window of 4 days ago or older, so scheduled runs only see days Google has finished processing. | No false alarms |
| Sampled rows in scripts | A script adds up query rows to get site totals and reports numbers that never match Search Console. | Gemini takes totals from Daily Metrics and uses the query and page datasets for comparisons, because those rows are sampled. | Totals that match |
| Control over each lookup | An automation with stored Search Console credentials can query any property, and no one sees what it asked for. | Gemini CLI asks you to confirm each Dataddo tool call, and you can set Always allow only for the tools you trust. | Lookups you approve |
| Token consumption | Gemini spends tokens exploring raw Google Search Console columns, sampling rows and retrying until it understands the data. | Field definitions and relations are supplied up front, so the model skips the exploration and goes straight to the answer. | Fewer tokens per answer |
| Answer quality | Prone to wrong joins, invented fields and numbers that are hard to check. | Answers grounded in defined fields. In the Dataddo benchmark, 78.5% of questions were answered correctly, against 65.5% with plain CSV files. | Better answers |
The Google Search Console datasets marketing teams use most, with their real field names. Gemini queries them by name through the Dataddo semantic layer.
| Dataset | Metrics and attributes |
Ask Gemini
|
|---|---|---|
| Daily Metrics |
|
Which Date in the last 16 months had the highest Impressions for our site? |
| Daily Metrics by Query |
|
Which Query values appeared for the first time this month with more than 100 Impressions? |
| Daily Metrics by Page |
|
Which URL values under /blog/ had Impressions this month but none last month? |
| Daily Metrics by Page and Query |
|
For each URL, which Query brings the most Clicks? |
| Daily Metrics by Country and Device |
|
Which Device has the most Impressions in each Country this quarter? |
| Sites Metadata |
|
Which Site URL properties fall under each Permission Level, listed for our config file? |
We asked 43 pre-registered questions of three real datasets - Google Search Console, HubSpot deals and Google Ads - and delivered the same rows to the same model in four ways. Only the data layer changed.
of questions answered correctly - the best of the four delivery methods tested
fewer confidently wrong answers - 14.0% of runs against 24.0% and 26.3%
more correct answers on messy CRM data - better on 8 questions, worse on none
Same model, prompts and rows in every condition; 759 graded runs, ground truth frozen before any run. On Google Search Console alone, documented CSV files did as well as Data to AI™ - the advantage comes from messier data such as HubSpot deals.
Read the benchmark →Add a Google Search Console source in Dataddo, for example Daily Metrics by Page, and attach it to a Google Gemini destination or AI Model. Put Dataddo in ~/.gemini/settings.json under mcpServers with httpUrl https://headless.dataddo.com/mcp-data, restart Gemini CLI and sign in with /mcp auth dataddo.
No. Dataddo extracts Search Console data into a flow, and Gemini CLI or Gemini Code Assist reads that flow through the Dataddo MCP server, like any other Dataddo source.
Sign-in opens a page in a web browser, so the machine running Gemini CLI needs a browser. Credentials are then stored in ~/.gemini/mcp-oauth-tokens.json and renewed automatically.
Query and page rows are sampled by Google, so their sums never match site totals. Ask Gemini to take totals from Daily Metrics or the country and device dataset.
Up to 16 months, the limit of the Google Search Console API. Load older days within that range once with a full data re-sync.
Either. ~/.gemini/settings.json makes the server available in every project, while .gemini/settings.json in a repository limits it to that project.
No. Gemini queries extracted data through the semantic layer by entity and field name. Dataddo never accepts SQL from the model and nothing is written back to Search Console.
