Generative search changed how users discover information, replacing the list of “10 blue links” with natural language responses. At the same time, it transformed the concept of visibility in search results.
Instead of solely monitoring SERP rankings, companies now must track mentions and citations in Large Language Models and explore in which context the brand shows up. The challenge is clear: How do you measure brand visibility in AI-driven search? Popular LLMs and Google don’t provide open data on AI-based searches and mentions, and manual monitoring across multiple AI services is neither scalable nor reliable.
But there is no need to worry. At DataForSEO, we developed a breakthrough data solution to prepare you for the era of generative search – LLM Mentions API. In this article, we’ll explore how to use this API to pull and track LLM mentions effortlessly. Moreover, we’ll demonstrate how generative search optimization (GEO) experts, tool developers, and brand agencies can transform LLM mentions data into actual insights.
Contents:
DataForSEO LLM Mentions API: a complete toolkit for AI mentions tracking
➤ How do these endpoints work if there is no open data on AI mentions?
➤ Search Mentions endpoint
➤ Aggregated Metrics endpoint
➤ Cross Aggregated Metrics endpoint
➤ Top Domains & Top Pages endpoints
Securing success in the generative search era with LLM Mentions API: 3 use cases
➤ Generative search optimization (GEO) experts
➤ Developers of AI visibility tracking tools
➤ Brand marketing and PR agencies
Conclusion
LLM Mentions API: a complete toolkit for tracking AI mentions
LLM Mentions API is a complete API suite for tracking mentions in LLM responses and retrieving relevant mention data. The API includes five endpoints – Search Mentions, Aggregated Metrics, Cross Aggregated Metrics, Top Domains, and Top Pages – each designed for collecting mention data grouped by specific parameters.
How do these endpoints work if there is no open data on AI mentions?
We use a sophisticated approach to address this limitation. First, we have formed an extensive database comprising millions of records of Google AI Overviews, sections of “people also ask” questions, and ChatGPT responses. This vast dataset is continuously expanded and regularly updated. Then, we structure the data and calculate all the relevant metrics, such as AI search volume, using our proprietary algorithms. After that, you can access the data instantly using LLM Mentions endpoints. This approach ensures you get complete, reliable AI mentions data whenever needed.
Now, let’s take a closer look at each of the LLM Mentions API endpoints.
1Search Mentions endpoint
The first endpoint, Search Mentions, provides detailed and structured mentions data for target keywords and domains. In particular, it fetches user queries and AI-generated snippets where the target is mentioned, complemented with mention metrics, such as total mention count and AI search volume.
The distinctive feature of the Search Mentions endpoint, as well as other endpoints of the LLM Mentions API, is the highly customizable task-setting process. First, you should specify the target array of up to 10 keyword or domain entities. For each entity in the target array, you can set custom search parameters, such as setting search scope for keyword or domain mentions, or excluding specific keywords from search results. This entity-based targeting approach allows you to narrow down search results precisely, making sure you receive only the most relevant mentions data.
Then, you can choose the platform type to get mentions data from a specific LLM platform (currently, either Google AI Overviews or ChatGPT). Besides, you can further customize the request by specifying location, language, and setting up ordering and filtering parameters. To manage the quantity of returned results, set the limit parameter to the preferred value.
Here is an example of a request to the Search Mentions endpoint:
[
{
"language_name": "English",
"location_code": 2840,
"target": [
{
"domain": "dataforseo.com",
"search_filter": "include"
},
{
"keyword": "dataforseo",
"search_scope": [
"answer"
]
}
],
"platform": "google",
"filters": [
[
"ai_search_volume",
">",
10
]
],
"order_by": [
"ai_search_volume,desc"
],
"offset": 0,
"limit": 2
}
]
The response will return results as follows:
{
"version": "0.1.20250922",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.2630 sec.",
"cost": 0.102,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "09261819-1535-0634-0000-57a9d2753804",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.1493 sec.",
"cost": 0.102,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"search",
"live"
],
"data": {
"api": "ai_optimization",
"function": "search",
"language_name": "English",
"location_code": 2840,
"target": [
{
"domain": "dataforseo.com",
"search_filter": "include"
},
{
"keyword": "dataforseo",
"search_scope": [
"answer"
]
}
],
"platform": "google",
"filters": [
[
"ai_search_volume",
">",
10
]
],
"order_by": [
"ai_search_volume,desc"
],
"offset": 0,
"limit": 2
},
"result": [
{
"total_count": 4,
"current_offset": 0,
"search_after_token": "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",
"items_count": 2,
"items": [
{
"platform": "google",
"location_code": 2840,
"language_code": "en",
"question": "google shopping api",
"answer": "The Google Shopping API refers primarily to the Content API for Shopping, which allows programmatic interaction with Google Merchant Center.This API is designed for merchants and developers to manage product data, inventory, and Merchant Center accounts more efficiently, especially for large or complex operations.\n\nKey functionalities of the Content API for Shopping include:\n\n- **Product Management:** Uploading, updating, and managing product data, including inventory, pricing, and product specifications.\n- **Account Management:** Managing Merchant Center accounts and sub-accounts, including linking accounts and handling account labels.\n- **Reporting and Insights:** Accessing performance metrics (clicks, impressions) and market insights like best-selling products and brands through the Reporting API.\n- **Integration with Google Ads:** Linking inventory to Google Ads and managing Shopping campaigns programmatically.\n\nOther APIs related to Google Shopping:\n\n- [**Google Ads API:** ](https://www.google.com/search?num=100&sca_esv=a9221caf88f3a082&hl=en&gl=US&glp=1&q=Google+Ads+API&sa=X&ved=2ahUKEwj5hPub2amPAxXhmbAFHX8gL_gQxccNegUI9wIQAw) Used for managing Google Ads campaigns, including Shopping ads, and linking them to Merchant Center data.\n- [**Merchant API:** ](https://www.google.com/search?num=100&sca_esv=a9221caf88f3a082&hl=en&gl=US&glp=1&q=Merchant+API&sa=X&ved=2ahUKEwj5hPub2amPAxXhmbAFHX8gL_gQxccNegUI9gIQAw) A newer API for managing various aspects of Google Merchant Center, offering similar functionalities to the Content API for Shopping.\n- [**Google Shopping Results API (from third-party providers like SerpApi):** ](https://www.google.com/search?num=100&sca_esv=a9221caf88f3a082&hl=en&gl=US&glp=1&q=Google+Shopping+Results+API+%28from+third-party+providers+like+SerpApi%29&sa=X&ved=2ahUKEwj5hPub2amPAxXhmbAFHX8gL_gQxccNegUI7wIQAw) These APIs focus on scraping and providing structured data from Google Shopping search results, including product details, pricing, and seller information.\n- [**Google Shopping Product Specs API (from third-party providers like DataForSEO):** ](https://www.google.com/search?num=100&sca_esv=a9221caf88f3a082&hl=en&gl=US&glp=1&q=Google+Shopping+Product+Specs+API+%28from+third-party+providers+like+DataForSEO%29&sa=X&ved=2ahUKEwj5hPub2amPAxXhmbAFHX8gL_gQxccNegUI_QIQAw) Offers detailed product specifications, including brand names, GTINs, and other technical parameters.\n- [**Google Shopping Reviews API (from third-party providers like DataForSEO):** ](https://www.google.com/search?num=100&sca_esv=a9221caf88f3a082&hl=en&gl=US&glp=1&q=Google+Shopping+Reviews+API+%28from+third-party+providers+like+DataForSEO%29&sa=X&ved=2ahUKEwj5hPub2amPAxXhmbAFHX8gL_gQxccNegUI_AIQAw) Provides access to customer reviews and ratings from Google Shopping listings.[[1]](https://dataforseo.com/apis/merchant-api-google-shopping#:~:text=Google%20Shopping%20Products%20API,integrate%20it%20into%20your%20solution.)\n\nThese APIs enable businesses to automate and integrate their product data and advertising efforts with Google's e-commerce ecosystem, enhancing efficiency and reach.",
"sources": [
{
"snippet": "Google Shopping Products API. Google Shopping Products API is a versatile engine for aggregating structured products data across G...",
"source_name": "DataForSEO",
"thumbnail": null,
"markdown": null,
"position": 1,
"title": "Merchant API: Google Shopping – Data on Products and Prices",
"domain": "dataforseo.com",
"url": "https://dataforseo.com/apis/merchant-api-google-shopping#:~:text=Google%20Shopping%20Products%20API,integrate%20it%20into%20your%20solution.",
"publication_date": null
}
],
"search_results": null,
"ai_search_volume": 880,
"monthly_searches": [
{
"year": 2025,
"month": 6,
"search_volume": 720
},
{
"year": 2025,
"month": 5,
"search_volume": 880
},
{
"year": 2025,
"month": 4,
"search_volume": 880
},
{
"year": 2025,
"month": 3,
"search_volume": 880
},
{
"year": 2025,
"month": 2,
"search_volume": 880
},
{
"year": 2025,
"month": 1,
"search_volume": 1300
},
{
"year": 2024,
"month": 12,
"search_volume": 720
},
{
"year": 2024,
"month": 11,
"search_volume": 720
},
{
"year": 2024,
"month": 10,
"search_volume": 720
},
{
"year": 2024,
"month": 9,
"search_volume": 590
},
{
"year": 2024,
"month": 8,
"search_volume": 590
},
{
"year": 2024,
"month": 7,
"search_volume": 590
}
]
},
{
"platform": "google",
"location_code": 2840,
"language_code": "en",
"question": "what is a callback ping",
"answer": "A callback ping is `a technical term for a server sending a brief, targeted HTTP or HTTPS request, or \"ping,\" to another server to establish a connection or trigger a response, often for real-time data transfer like performance marketing event data`.In the context of mobile apps, it sends structured data such as event type or device ID to an analytics platform.In other contexts, it can mean a customer service request, a remote access request, or a notification in a work-related or video game context.[[1]](https://www.youtube.com/shorts/JW9Za7pgRaY)[[2]](https://www.adjust.com/glossary/callback/#:~:text=The%20definition%20of%20callback,for%20performance-driven%20mobile%20marketing.)[[3]](https://docs.pingidentity.com/pingoneaic/latest/am-authentication/callbacks-supported.html#:~:text=For%20an%20example%20using%20callbacks,Core%20authentication%20attributes%20Interactive%20callbacks)[[4]](https://dataforseo.com/help-center/pingbacks-postbacks-with-dataforseo-api#:~:text=Commonly%2C%20pingback%20is%20known%20as%20a%20callback,actually%20sends%20the%20results%20to%20your%20server.)[[5]](https://www.remoteutilities.com/support/docs/callback-connection/#:~:text=The%20primary%20concept%20of%20the,or%20automatically%20at%20regular%20intervals.)[[6]](https://www.five9.com/faq/what-is-a-callback#:~:text=A%20callback%20is%20a%20request,on%20hold%20on%20the%20phone.)[[7]](https://www.audiokinetic.com/courses/wwise301/?source=wwise301&id=Callbacks#:~:text=As%20you%20might%20deduce%20from%20the%20name%2C,hit%20occurs%20in%20the%20orchestral%20music%20ensemble.)\n\nKey Characteristics[[1]](https://www.adjust.com/glossary/callback/#:~:text=The%20definition%20of%20callback,for%20performance-driven%20mobile%20marketing.)\n\n- **Server-to-Server Communication**:A callback ping is fundamentally a signal sent from one server to another, typically for data exchange.\n- **Event-Driven**:It's often triggered by a user's interaction with an app, like an ad engagement or an in-app action, to send immediate, raw event data for analytics.\n- **Data Transfer**:The ping carries structured data, such as advertising IDs, event types, and timestamps, to a different system for analysis or other purposes.\n- **Real-Time Functionality**:Callbacks enable real-time data delivery and help mobile marketers accurately track performance.\n\nExamples\n\n- **Mobile Marketing**:When a user clicks an ad, the ad tracker sends a callback ping with specific details to an analytics provider to report the engagement.[[1]](https://www.adjust.com/glossary/callback/#:~:text=The%20definition%20of%20callback,for%20performance-driven%20mobile%20marketing.)\n- **Remote Support**:A host can \"ping\" a viewer to request a connection, which the viewer then accepts to grant remote access, according to Remote Utilities.[[1]](https://www.remoteutilities.com/support/docs/callback-connection/#:~:text=The%20primary%20concept%20of%20the,or%20automatically%20at%20regular%20intervals.)\n- **Gaming/Fiction**:In the show Murder Drones, a character uses a \"callback ping\" to establish a connection with another server, potentially to regain control or transfer information.[[1]](https://www.youtube.com/shorts/JW9Za7pgRaY)",
"sources": [
{
"snippet": "Aug 13, 2024 — next shot of sing. we get however seems pretty clear that it is an outpost. 3 in this shot where she seems to be compl...",
"source_name": "YouTube ·",
"thumbnail": null,
"markdown": null,
"position": 1,
"title": "So about that Callback Ping... #shorts #murderdrones #theory #animation ...",
"domain": "www.youtube.com",
"url": "https://www.youtube.com/shorts/JW9Za7pgRaY",
"publication_date": null
},
{
"snippet": "The definition of callback. A callback, also known as a postback, is a ping (an http or https request) made between one server and...",
"source_name": "Adjust",
"thumbnail": null,
"markdown": null,
"position": 2,
"title": "What is a callback? - Adjust",
"domain": "www.adjust.com",
"url": "https://www.adjust.com/glossary/callback/#:~:text=The%20definition%20of%20callback,for%20performance-driven%20mobile%20marketing.",
"publication_date": null
},
{
"snippet": "For an example using callbacks to authenticate, refer to Return callback information to Advanced Identity Cloud. Advanced Identity...",
"source_name": "Ping Identity",
"thumbnail": null,
"markdown": null,
"position": 3,
"title": "Supported callbacks | PingOne Advanced Identity Cloud",
"domain": "docs.pingidentity.com",
"url": "https://docs.pingidentity.com/pingoneaic/latest/am-authentication/callbacks-supported.html#:~:text=For%20an%20example%20using%20callbacks,Core%20authentication%20attributes%20Interactive%20callbacks",
"publication_date": null
},
{
"snippet": "Jul 22, 2021 — Commonly, pingback is known as a callback function, and postback is referred to as a webhook or HTTP push. With DataFo...",
"source_name": "DataForSEO",
"thumbnail": null,
"markdown": null,
"position": 4,
"title": "What are pingbacks & postbacks and how to use them with DataForSEO API? – DataForSEO",
"domain": "dataforseo.com",
"url": "https://dataforseo.com/help-center/pingbacks-postbacks-with-dataforseo-api#:~:text=Commonly%2C%20pingback%20is%20known%20as%20a%20callback,actually%20sends%20the%20results%20to%20your%20server.",
"publication_date": null
},
{
"snippet": "The primary concept of the callback connection type is that the Host \"pings\" the Viewer and offers remote access. To initiate a re...",
"source_name": "Remote Utilities",
"thumbnail": null,
"markdown": null,
"position": 5,
"title": "Callback Connection | Remote Utilities",
"domain": "www.remoteutilities.com",
"url": "https://www.remoteutilities.com/support/docs/callback-connection/#:~:text=The%20primary%20concept%20of%20the,or%20automatically%20at%20regular%20intervals.",
"publication_date": null
},
{
"snippet": "A callback is a request by a customer to have a customer service agent call them back. It is often part of a queue callback featur...",
"source_name": "Five9",
"thumbnail": null,
"markdown": null,
"position": 6,
"title": "What is a Callback in a Call Center - Five9",
"domain": "www.five9.com",
"url": "https://www.five9.com/faq/what-is-a-callback#:~:text=A%20callback%20is%20a%20request,on%20hold%20on%20the%20phone.",
"publication_date": null
},
{
"snippet": "As you might deduce from the name, callbacks allow you to get a call back from an Event you have posted. It is a way for the game ...",
"source_name": "Audiokinetic",
"thumbnail": null,
"markdown": null,
"position": 7,
"title": "Wwise Unity Integration - Callbacks",
"domain": "www.audiokinetic.com",
"url": "https://www.audiokinetic.com/courses/wwise301/?source=wwise301&id=Callbacks#:~:text=As%20you%20might%20deduce%20from%20the%20name%2C,hit%20occurs%20in%20the%20orchestral%20music%20ensemble.",
"publication_date": null
}
],
"search_results": null,
"ai_search_volume": 260,
"monthly_searches": [
{
"year": 2025,
"month": 6,
"search_volume": 90
},
{
"year": 2025,
"month": 5,
"search_volume": 140
},
{
"year": 2025,
"month": 4,
"search_volume": 90
},
{
"year": 2025,
"month": 3,
"search_volume": 90
},
{
"year": 2025,
"month": 2,
"search_volume": 170
},
{
"year": 2025,
"month": 1,
"search_volume": 170
},
{
"year": 2024,
"month": 12,
"search_volume": 260
},
{
"year": 2024,
"month": 11,
"search_volume": 390
},
{
"year": 2024,
"month": 10,
"search_volume": 590
},
{
"year": 2024,
"month": 9,
"search_volume": 590
},
{
"year": 2024,
"month": 8,
"search_volume": 390
},
{
"year": 2024,
"month": 7,
"search_volume": null
}
]
}
]
}
]
}
]
}
In the items array of results, you will receive objects that contain mention data for the relevant target. These objects include a user question and an LLM answer in which the target was mentioned, a list of sources quoted by AI in the response, and all websites the model retrieved when looking up the information. Additionally, you get the current and monthly ai_search_volume values for the target keyword or domain. This data is especially beneficial for exploring how and in which context keywords and domains are mentioned in AI responses.
2Aggregated Metrics endpoint
The next endpoint, Aggregated Metrics, gives you a consolidated overview of key mention metrics across different dimensions, such as location, language, AI platform, and source domains.
The task-setting process for this endpoint is similar to that for the Mentions endpoint. Set the target array with the target keyword or domain entities and specify the location and language. Then, select the platform to retrieve the results from and define the filtering parameters.
The request may be structured like this:
[
{
"language_code": "es",
"location_code": 2840,
"platform": "google",
"target": [
{
"domain": "en.wikipedia.org",
"search_filter": "exclude"
},
{
"keyword": "bmw",
"search_scope": [
"answer"
]
}
],
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"internal_list_limit": 3
}
]
The response will return the following results:
{
"version": "0.1.20250922",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.6480 sec.",
"cost": 0.101,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "09261821-1535-0635-0000-5864ad3ff1e0",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.5832 sec.",
"cost": 0.101,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"aggregated_metrics",
"live"
],
"data": {
"api": "ai_optimization",
"function": "aggregated_metrics",
"language_code": "es",
"location_code": 2840,
"platform": "google",
"target": [
{
"domain": "en.wikipedia.org",
"search_filter": "exclude"
},
{
"keyword": "bmw",
"search_scope": [
"answer"
]
}
],
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"internal_list_limit": 3
},
"result": [
{
"total": {
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 483,
"ai_search_volume": 2432100,
"impressions": 1174704300
}
],
"language": [
{
"type": "group_element",
"key": "es",
"mentions": 483,
"ai_search_volume": 2432100,
"impressions": 1174704300
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 483,
"ai_search_volume": 2432100,
"impressions": 1174704300
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 318,
"ai_search_volume": 1513700,
"impressions": 481356600
},
{
"type": "group_element",
"key": "www.caranddriver.com",
"mentions": 103,
"ai_search_volume": 510800,
"impressions": 52612400
},
{
"type": "group_element",
"key": "cars.usnews.com",
"mentions": 79,
"ai_search_volume": 710300,
"impressions": 56113700
}
],
"search_results_domain": null
},
"items": null
}
]
}
]
}
The API returns a response with the aggregated mention metrics grouped by specified location, language, and LLM. The metrics include current ai_search_volume and impressions values, as well as total mentions count. Besides, the mention data is additionally grouped by all relevant domains found in the AI responses. With this endpoint, you can quickly assess the overall traffic potential of your target mentions and identify which domains consistently appear in related AI responses.
3Cross Aggregated Metrics endpoint
The Cross Aggregated Metrics endpoint also provides consolidated mention metrics, but with a key difference. In this endpoint, you can specify multiple target mentions to get the data for. It features a separate targets array, where you can set up multiple target domains or keywords grouped under aggregation keys. For each individual target, you can set up a search scope and filtering parameters. The rest of the request parameters, such as location, language, and initial_dataset_filters, are specified similarly to the way it works in the Aggregated Metrics endpoint. This setup allows you to compare mention metrics across multiple targets within a single request.
For instance, you want to compare how frequently different phone models like iPhone, Samsung, Google Pixel, and Xiaomi appear alongside the keyword “best camera” in AI responses. To do that, you can send the following request to the Cross Aggregated Metrics endpoint:
[
{
"language_code": "en",
"location_code": 2840,
"platform": "google",
"targets": [
{
"aggregation_key": "iphone",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "iphone",
"search_filter": "include"
}
]
},
{
"aggregation_key": "samsung",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "samsung",
"search_filter": "include"
}
]
},
{
"aggregation_key": "google pixel",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "google pixel",
"search_filter": "include"
}
]
},
{
"aggregation_key": "xiaomi",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "xiaomi",
"search_filter": "include"
}
]
}
],
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"internal_list_limit": 2
}
]
As a result, you will get a response like this one below:
{
"version": "0.1.20250922",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.8062 sec.",
"cost": 0.101,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "09261823-1535-0636-0000-c71a3f6aa09a",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.7761 sec.",
"cost": 0.101,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"cross_aggregated_metrics",
"live"
],
"data": {
"api": "ai_optimization",
"function": "cross_aggregated_metrics",
"language_code": "en",
"location_code": 2840,
"platform": "google",
"targets": [
{
"aggregation_key": "iphone",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "iphone",
"search_filter": "include"
}
]
},
{
"aggregation_key": "samsung",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "samsung",
"search_filter": "include"
}
]
},
{
"aggregation_key": "google pixel",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "google pixel",
"search_filter": "include"
}
]
},
{
"aggregation_key": "xiaomi",
"target": [
{
"keyword": "best camera",
"search_filter": "include"
},
{
"keyword": "xiaomi",
"search_filter": "include"
}
]
}
],
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"internal_list_limit": 2
},
"result": [
{
"total": {
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 12593,
"ai_search_volume": 3846850,
"impressions": 48443382050
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 12593,
"ai_search_volume": 3846850,
"impressions": 48443382050
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 12593,
"ai_search_volume": 3846850,
"impressions": 48443382050
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 30898,
"ai_search_volume": 2961270,
"impressions": 91497320460
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 8626,
"ai_search_volume": 1282340,
"impressions": 11061464840
}
],
"search_results_domain": null
},
"items": [
{
"key": "iphone",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 10084,
"ai_search_volume": 3248310,
"impressions": 32755958040
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 10084,
"ai_search_volume": 3248310,
"impressions": 32755958040
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 10084,
"ai_search_volume": 3248310,
"impressions": 32755958040
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 8278,
"ai_search_volume": 2523600,
"impressions": 20890360800
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 4087,
"ai_search_volume": 1085400,
"impressions": 4436029800
}
],
"search_results_domain": null
},
{
"key": "samsung",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 3571,
"ai_search_volume": 916970,
"impressions": 3274499870
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 3571,
"ai_search_volume": 916970,
"impressions": 3274499870
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 3571,
"ai_search_volume": 916970,
"impressions": 3274499870
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 2844,
"ai_search_volume": 588640,
"impressions": 1674092160
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 1572,
"ai_search_volume": 286530,
"impressions": 450425160
}
],
"search_results_domain": null
},
{
"key": "google pixel",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 1272,
"ai_search_volume": 450060,
"impressions": 572476320
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 1272,
"ai_search_volume": 450060,
"impressions": 572476320
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 1272,
"ai_search_volume": 450060,
"impressions": 572476320
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 949,
"ai_search_volume": 265120,
"impressions": 251598880
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 556,
"ai_search_volume": 164480,
"impressions": 91450880
}
],
"search_results_domain": null
},
{
"key": "xiaomi",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 450,
"ai_search_volume": 232060,
"impressions": 104427000
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 450,
"ai_search_volume": 232060,
"impressions": 104427000
}
],
"platform": [
{
"type": "group_element",
"key": "google",
"mentions": 450,
"ai_search_volume": 232060,
"impressions": 104427000
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.youtube.com",
"mentions": 295,
"ai_search_volume": 102230,
"impressions": 30157850
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 166,
"ai_search_volume": 71460,
"impressions": 11862360
}
],
"search_results_domain": null
}
]
}
]
}
]
}
In the response, locate the items array. It contains mention metrics for each individual target, grouped under aggregation keys. Thus, you can clearly see how often different phone models are mentioned in AI responses alongside the keyword “best camera”. Besides, the response features a separate total object with combined metrics of all targets.
4Top Domains & Top Pages endpoints
Finally, the last two endpoints, Top Domains and Top Pages, help you identify which domains and pages appear most frequently in AI responses containing your specific keyword or domain mention. The task-setting process for these endpoints is essentially the same as in the Aggregated Metrics endpoint, but here you can specify the additional links_scope parameter. This parameter allows you to select which types of links from AI responses will be used to extract domain or page data. You can choose between two link types:
sources– links quoted directly in the AI response;search_results– all the links the LLM found during web search.
Here is an example request to the Top Domains endpoint:
[
{
"language_code": "en",
"location_code": 2840,
"platform": "chat_gpt",
"target": [
{
"keyword": "bmw",
"search_scope": [
"answer"
]
},
{
"keyword": "auto",
"search_scope": [
"question"
],
"match_type": "partial_match"
}
],
"links_scope": "sources",
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"items_list_limit": 2,
"internal_list_limit": 2
}
]
The response will return results as follows:
{
"version": "0.1.20250922",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.7419 sec.",
"cost": 0.101,
"tasks_count": 1,
"tasks_error": 0,
"tasks": [
{
"id": "09261824-1535-0637-0000-3fd6c87eaa56",
"status_code": 20000,
"status_message": "Ok.",
"time": "0.6960 sec.",
"cost": 0.101,
"result_count": 1,
"path": [
"v3",
"ai_optimization",
"llm_mentions",
"top_domains",
"live"
],
"data": {
"api": "ai_optimization",
"function": "top_domains",
"language_code": "en",
"location_code": 2840,
"platform": "chat_gpt",
"target": [
{
"keyword": "bmw",
"search_scope": [
"answer"
]
},
{
"keyword": "auto",
"search_scope": [
"question"
],
"match_type": "partial_match"
}
],
"links_scope": "sources",
"initial_dataset_filters": [
[
"ai_search_volume",
">",
10
]
],
"items_list_limit": 2,
"internal_list_limit": 2
},
"result": [
{
"total": {
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 62,
"ai_search_volume": 3913,
"impressions": 242606
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 62,
"ai_search_volume": 3913,
"impressions": 242606
}
],
"platform": [
{
"type": "group_element",
"key": "chat_gpt",
"mentions": 62,
"ai_search_volume": 3913,
"impressions": 242606
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 30,
"ai_search_volume": 1485,
"impressions": 44550
},
{
"type": "group_element",
"key": "en.wikipedia.org",
"mentions": 25,
"ai_search_volume": 1142,
"impressions": 28550
}
],
"search_results_domain": [
{
"type": "group_element",
"key": "www.indeed.com",
"mentions": 4,
"ai_search_volume": 697,
"impressions": 2788
},
{
"type": "group_element",
"key": "www.autonews.com",
"mentions": 3,
"ai_search_volume": 380,
"impressions": 1140
}
]
},
"items": [
{
"key": "www.reddit.com",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 30,
"ai_search_volume": 1485,
"impressions": 44550
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 30,
"ai_search_volume": 1485,
"impressions": 44550
}
],
"platform": [
{
"type": "group_element",
"key": "chat_gpt",
"mentions": 30,
"ai_search_volume": 1485,
"impressions": 44550
}
],
"sources_domain": [
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 30,
"ai_search_volume": 1485,
"impressions": 44550
},
{
"type": "group_element",
"key": "en.wikipedia.org",
"mentions": 9,
"ai_search_volume": 292,
"impressions": 2628
}
],
"search_results_domain": [
{
"type": "group_element",
"key": "blog.torque360.co",
"mentions": 3,
"ai_search_volume": 89,
"impressions": 267
},
{
"type": "group_element",
"key": "www.collegevine.com",
"mentions": 2,
"ai_search_volume": 88,
"impressions": 176
}
]
},
{
"key": "en.wikipedia.org",
"location": [
{
"type": "group_element",
"key": "2840",
"mentions": 25,
"ai_search_volume": 1142,
"impressions": 28550
}
],
"language": [
{
"type": "group_element",
"key": "en",
"mentions": 25,
"ai_search_volume": 1142,
"impressions": 28550
}
],
"platform": [
{
"type": "group_element",
"key": "chat_gpt",
"mentions": 25,
"ai_search_volume": 1142,
"impressions": 28550
}
],
"sources_domain": [
{
"type": "group_element",
"key": "en.wikipedia.org",
"mentions": 25,
"ai_search_volume": 1142,
"impressions": 28550
},
{
"type": "group_element",
"key": "www.reddit.com",
"mentions": 10,
"ai_search_volume": 315,
"impressions": 3150
}
],
"search_results_domain": [
{
"type": "group_element",
"key": "detroitautoshow.com",
"mentions": 1,
"ai_search_volume": 40,
"impressions": 40
},
{
"type": "group_element",
"key": "www.collegevine.com",
"mentions": 2,
"ai_search_volume": 88,
"impressions": 176
}
]
}
]
}
]
}
]
}
In the total object of the result, you will find aggregated mention metrics of all the top domains (or pages) found for the specified target. The detailed metrics of each found domain or page are located in separate objects of the items array. Each object features the key field that contains the top domain name or top page URL. In addition, for every found domain or page, you will receive the current ai_search_volume and the total number of mentions. With this data, you can easily explore which domains or pages have the best AI visibility in your target market.
To explore all the LLM Mentions endpoints in more detail, especially the peculiarities of the task-setting process, read our comprehensive guide at the Help Center.
The other important aspect to consider about this API is its pricing system. It features a monthly minimum commitment of $100, similar to the Backlinks API. We introduced this commitment to cover the cost of data retrieval and scale up our existing mentions database. For each of the LLM mentions endpoints, the usage price is $0.01 for setting a task and $0.001 per item in the result. This means that if you get a response with five separate data items, you will pay $0,015 in total. You can limit the number of items in the result with the limit or items_list_limit parameter. For more information about data costs, check out our Pricing page.
Now you understand what makes the LLM Mentions API an advanced toolkit for collecting precise and structured LLM mention data. However, what are the best possible ways to leverage this data, and who can benefit from it the most? Let’s explore three real-world use cases that demonstrate the practical applications.
Securing success in the generative search era with LLM Mentions API: 3 use cases
The DataForSEO LLM Mentions API removes the curtain of uncertainty in tracking mentions in fluid conversational AI responses. It provides diverse digital industry professionals with vast opportunities to excel in the generative search era. Especially, the LLM Mentions API is the best solution for:
- Generative search optimization (GEO) experts, who want to enhance AI visibility of their customers’ businesses;
- Software developers who want to build a new generation of tools for tracking generative search performance;
- Brand agencies that want to assess and improve brands’ perception and share of voice in LLM searches.
Let’s see how they can utilize the LLM mentions insights to achieve their goals most effectively.
1Generative search optimization (GEO) experts
For the new generation of GEO experts, who guide the companies through the unexplored land of generative search, access to LLM mentions data is a cornerstone of success. Without such data, it’s neither possible to evaluate a business’s AI visibility nor find ways to improve it.
However, with the LLM Mentions API, GEO experts can effortlessly perform the most complex generative search optimization tasks. In particular, they can do the following:
➤ Explore the company’s competitors in generative LLM search by key mentions. With the help of the Search Mentions, Top Domains, and Top Pages endpoints, you can quickly assess what companies already dominate the AI results by key industry mentions. For instance, the Search Mentions endpoint can provide a clear picture of the AI-generated results, where competitors rank for industry-specific keywords. Besides, you can analyze the sources array of each AI response. It contains URLs of all websites mentioned by AI and respective text snippets included in the response, providing you an understanding of the typesof sources the LLM considers authoritative.
With the Top Domains and Top Pages endpoints, you can conduct broader competitor research and assess competitors’ overall share of voice in the LLMs. By examining the mentions_count, ai_search_volume, and impressions metrics for top websites, it is possible to determine which competitors consistently capture the most AI attention and traffic in your industry.
➤ Analyze traffic potential and impression rate of target mentions in LLMs. Understanding how much traffic mentions can generate for a website is as important as tracking its AI visibility. Suppose your website is frequently mentioned in LLMs, but the share of AI traffic remains low. That’s the signal to explore new keyword mentions with better traffic potential. For this purpose, you can use Aggregated Metrics and Cross Aggregated Metrics endpoints.
By using the Aggregated metrics endpoint, you can quickly find out how much AI traffic your target keywords can potentially bring to the website. This endpoint provides current ai_search_volume and impressions values for the keyword, grouped by multiple dimensions, such as location, language, and AI platform. You can run this analysis for various keywords and figure out which ones appear prominently in LLMs and attract significant traffic. To facilitate the research even more, you can use the Cross Aggregated Metrics endpoint. In this endpoint, you can specify multiple target keywords and get search volume and impressions data for them in a single request. By comparing the keywords’ metrics side-by-side, you can quickly identify the ones with the highest potential.
➤ Monitor changes in the company’s AI visibility over time. Consistent tracking of a website’s AI visibility and key mention metrics is crucial for measuring the effectiveness of generative search optimization strategies. By regularly pulling mentions data with the LLM Mentions API, you can precisely monitor changes in mention frequency, AI traffic volume, and establish KPIs to track the optimization progress. Based on the monitoring data, you can create comprehensive reports that clearly demonstrate the changes in the website’s AI presence and possible ways to fine-tune your optimization strategy.
As you can see, with the help of the LLM Mentions API, GEO experts can easily perform the most complex AI visibility optimization tasks in a data-driven way.
2Developers of AI visibility tracking tools
For developers of search optimization tools, access to the actual and structured LLM mentions data is vital. With this data, they can incorporate advanced AI visibility tracking features into existing tools or build next-generation GEO solutions from scratch.
Let’s review the most practical and valuable examples of new features that you can build with the LLM Mentions API.
1. LLM queries and responses explorer. First, you can build a customizable LLM queries and responses explorer, like the one in the Ahrefs Brand Radar. By using such an explorer, users can search for AI responses and queries containing specific keyword mentions.
Image source: Ahrefs Blog
Most of the elements present in the explorer can be built with the data from the Search Mentions endpoint:
- For the “AI Overview” element, which displays the generated answer and list of quoted links, you can use data from the
answerfield andsourcesarray of API response. - The respective AI Overview queries (“Keyword” element) can be fetched from the
questionfields in the response. - Fields
total_countandai_search_volumecan be used to display the total mentions count and the traffic volume, respectively.
2. AI visibility overview dashboard. The next feature you can build is an interactive dashboard for tracking changes in your and your competitors’ AI visibility over time. This dashboard can include bar charts and a graph to display the current AI share of voice of tracked websites and historical changes in key mention metrics.
The Cross Aggregated Metrics endpoint is perfect for building this dashboard. With this endpoint, you can track multiple domains simultaneously and receive mention metrics for each. As a result, it is possible to create graphs and charts that compare metrics of domains side by side.
In particular, you can do the following:
- Pull
ai_search_volume,impressions, andtotal_countof mentions for different domains for cross-comparison. - Consistently collect and track mentions data to visualise historical changes.
- Retrieve data for different AI platforms for a comprehensive overview of AI visibility
Overall, with the Cross Aggregated Metrics endpoint, you can design a functional and user-friendly dashboard that will track and visualise the slightest changes in websites’ AI visibility.
3. Top cited domains and pages lists. Using data from the Top Domains and Top Pages endpoints, you can create lists showing the most frequently mentioned domains and pages in LLMs. This table enables users to investigate which domains or pages have the most significant visibility in AI results by target queries.
Aside from exploring top domains and page URLs, users can also see how many impressions, mentions, and AI search volume each domain/page has in LLM responses. All of these metrics can be pulled straight from the Top Domains/Top Pages endpoints’ responses and displayed in the table. Additionally, users can compare the top domains with selected industry competitors by the number of mentions or other metrics.
These tables can be handy for assessing how fierce competition is in AI searches within a specific industry. Besides, you can evaluate whether it is worth targeting specific mentions in AI responses, considering the competition level.
Overall, the LLM Mentions API allows developers to design powerful AI visibility tracking features or build standalone GEO tools, providing immense value to customers.
3Brand marketing and PR agencies
The shift to generative search influences brand marketing and PR, too. Now, AI emerges as a separate channel for brand monitoring because users can consider AI-generated responses as authoritative sources. That means the way LLM mentions your brand in the response, the context AI gives to the mention can determine users’ actions towards your brand. In this situation, tracking the AI responses for brand mentions and analyzing their context and sentiment is crucial for brand strategies.
Fortunately, with the LLM Mentions API, brand agencies can instantly access all required AI mentions data. For example, by using the Search Mentions endpoint, they can collect numerous snippets of LLM responses for specific brand or product mentions. Then, the agency can perform a deep sentiment analysis of relevant responses. This analysis will allow us to determine the sentiment the brand receives the most, how frequently, and in which context the brand appears in AI responses. With detailed sentiment insights, the brand agency can develop data-driven recommendations for improving brand positioning and the company’s value proposition in content marketing and communications.
Beyond sentiment analysis, brand agencies must also focus on shaping how AI models understand and represent their clients’ brands. This can be done by supplying LLMs with relevant information that positions brands favorably through authoritative channels. In this context, the Search Mentions and Top Domains endpoints can help to explore such channels quickly. For instance, in the Search Mentions endpoint, marketers can explore the sources arrays in relevant responses to see which types of sources are favored by AI. Suppose there are websites such as Quora, Reddit, or thematic forums among the sources. In that case, it may be worth considering increasing the brand presence on such resources through communication or content campaigns.
Similarly, marketers can use the Top Domains endpoint to see if there are any relevant channels among the top domains by mentions. They can also estimate the traffic potential from relevant domains by considering ai_search_volume, mentions, and impressions metrics. As a result, a brand marketing agency can map out the best potential channels for an AI brand positioning strategy to make it successful.
In short, these three use cases demonstrate that the LLM Mentions API is a versatile solution that helps many professionals remain confident in the era of generative search. From building advanced mention tracking tools to shaping the brand image for AI, the LLM Mentions API serves as a reliable and quality data foundation.
Conclusion
The generative search era surprised many of us, bringing uncertainty with AI visibility tracking and optimization issues. But the times of uncertainty are over. With the DataForSEO LLM mentions API, AI visibility tracking is no longer guesswork, but a data-driven process. This API transforms conversational responses into structured mentions data that can help various professionals who want to secure success in generative search.
In particular, the power of the LLM Mentions API allows you to:
- Carry out complex generative search optimization tasks, like competitor research, mention monitoring, and AI keyword research;
- Develop advanced AI visibility tracking tools;
- Collect clear and structured AI response data for brand marketing and sentiment analysis.
Register at DataForSEO now and step into the generative search era with the best data solution on the market!