DataForSEO LLM Mentions API Review: Tracking ChatGPT and Gemini Visibility

What are DataForSEO's AI visibility APIs?

DataForSEO groups them under its AI Optimization API. The three that matter for visibility work differently:

The distinction between the last two is important. DataForSEO's own help centre explains that LLM Responses shows what the underlying model produces through its API, not what appears in the platform's interface. LLM Scraper reads the interface users actually see. For visibility tracking, the interface is usually what you want.

What does the LLM Mentions API offer?

According to DataForSEO's product page and launch notes:

• Search Mentions: find where a keyword or domain appears in AI answers, with the full question and answer, cited and non-cited sources, AI search volume and monthly searches.
• Target Metrics: how often a domain or keyword is mentioned, grouped by location, language, platform and domain.
• Cross Aggregated Metrics: several brands or competitors side by side in one request.
• Top Domains, Top Pages and Top Brands: the sites, pages and brands AI answers mention most for a topic.
• Historical endpoints: month-by-month mention counts back to 2025, changes between two dates, and new and lost mentions.
• Lite endpoints: the same core counts, trimmed for speed.

DataForSEO says responses return within about two seconds. The prompts come from DataForSEO's own database, so you are measuring your share of a large sample of real-world questions, not the specific prompts your buyers use. That is useful for market-level benchmarking, and less useful for tracking a short list of high-value questions.

Does it track Gemini?

Not in LLM Mentions. DataForSEO describes its data as covering Google's AI Overviews and ChatGPT; at launch it said the ChatGPT data was US-only and from GPT-5. To see Gemini answers, use:

• LLM Scraper for Gemini, which scrapes the Gemini web version for your prompt, location and language, and returns the text, sources and item types such as tables and images.
• LLM Responses for Gemini, which calls Google's developer API. That lets you choose model versions and parameters, but the answer may differ from what Gemini users see.

Neither builds a history for you. You store each day's results and calculate visibility yourself.

What does tracking your own prompts cost?

LLM Scraper is priced per result page. Tracking 100 prompts daily on both ChatGPT and Gemini means about 6,000 results a month:

That is far cheaper than any dashboard. The real cost is building and running the system around it: storing results, detecting brand mentions and positions, handling errors and changes to the response format, and presenting the data. Compare it with ready-made trackers such as SE Visible, which sells 200 prompts on five engines for $99 a month.

How to use the DataForSEO API for SEO and GEO audits

DataForSEO's range of APIs makes it a practical base for an audit that covers both search and AI answers. A workable sequence:

1. Crawl the site with the On-Page API. The basic crawl costs $0.00015 per page, or $0.15 per 1,000 pages. It reports internal links, duplicates, page speed and non-indexable pages. Loading resources and rendering JavaScript cost more, but show what crawlers without JavaScript miss.
2. Check crawler access separately. Read robots.txt and your CDN or firewall rules for the AI crawlers you care about. An API crawl does not tell you whether GPTBot or other AI user agents are blocked at the edge.
3. Pull search data for your target keywords with the SERP API, including whether AI Overviews appear, and the Google AI Mode SERP API for AI Mode answers.
4. Benchmark AI mentions with LLM Mentions: Cross Aggregated Metrics for you and competitors, and Top Pages to see which of your pages AI answers cite.
5. Test your own prompts with LLM Scraper on ChatGPT and Gemini, 20 to 50 prompts covering category, comparison and brand questions.
6. Join the results: pages that rank but are never cited, pages cited without ranking, and technical issues on pages that should be cited.

DataForSEO also offers an MCP server that connects its APIs to Claude Desktop, Claude Code, Cursor and ChatGPT. With it, an assistant can call the APIs and draft the audit in minutes. Two cautions apply. Every call costs money, so set limits on crawls and requests before letting an assistant run. And check the numbers in the draft against the raw API output, because summaries can misstate counts.

What are the limits?

• Engine coverage: LLM Mentions covers AI Overviews and US ChatGPT only; Gemini, Perplexity and Claude need other APIs.
• Database prompts: LLM Mentions measures DataForSEO's prompt sample, not your chosen prompts.
• Monthly commitment: the $100 top-up applies each month you use LLM Mentions.
• Build effort: raw data needs engineering to become a tracker or report.
• Single runs: each scraper result is one answer; you decide how many runs per prompt to average.

Who is it for?

• Good fit: SEO software builders, agencies with developers, and data teams that want AI visibility in their own dashboards.
• Reasonable fit: technical SEOs who use the MCP server for one-off audits.
• Poor fit: marketers who want a dashboard without writing code.

Other ways to track the same engines are covered in the VisiGEO review, the Gauge review and the SpyFu ChatGPT mention tracking review.

What might an agency build look like?

Consider an illustrative five-person agency, Raskoway Search, that wanted AI visibility in its existing client reports. It used LLM Scraper to run 40 prompts per client daily on ChatGPT and Gemini, storing results in its own database, at a few dollars per client a month. Once a month, it pulled Cross Aggregated Metrics from LLM Mentions to benchmark each client against three competitors in AI Overviews. Building the pipeline took one developer about two weeks. The agency judged that worthwhile because it served 18 clients; for three clients, it would have bought a tracker.

What do people ask about the DataForSEO LLM Mentions API?

How much does it cost?

$0.10 per request plus $0.001 per row, with at least $100 topped up each month.

Which AI platforms does it cover?

Google AI Overviews and ChatGPT. Gemini answers come from the LLM Scraper or LLM Responses APIs.

Can I track my own prompts?

Not with LLM Mentions, which searches DataForSEO's database. Use LLM Scraper for your own prompts.

Can I use it with Claude?

Yes. DataForSEO's MCP server connects its APIs to Claude Desktop and Claude Code.