Skip to main content

Overview


LLM Responses

Query major AI models and capture their responses.

Supported LLMs

Options

--keyword is sent as user_prompt to the API. The model field is model_name, and each platform only accepts its own models — you cannot pass an OpenAI model name to Claude, Gemini, or Perplexity. List valid models with naive aeo llm-responses models --platform <chatgpt|claude|gemini|perplexity>.When --model is omitted, a valid per-platform default is used automatically: chatgptgpt-4o-mini, claudeclaude-sonnet-4-5, geminigemini-2.5-flash, perplexitysonar. Perplexity only supports Live mode and only the sonar, sonar-pro, and sonar-reasoning-pro models. Do not send location/language to llm-responses.

LLM Scraper

Scrape ChatGPT Search results including citations and source links.

Options


LLM Mentions

Track how frequently brands and domains appear in AI responses.
DataForSEO requires a target field that is an array of objects, each { "keyword": "..." } or { "domain": "..." }. A top-level keyword/domain string is invalid. The CLI builds this array for you: search, top-domains, and top-pages take --keyword (→ target: [{ keyword }]); aggregated-metrics and cross-aggregated-metrics take --target <domain> (→ target: [{ domain }]).

Endpoints

Options

There is no date-range flag on llm-mentions. The CLI declares --keyword, --target, --location-code and --language-code and nothing else; anything narrower has to be filtered from the result.

AI Keywords

Get search volume estimates for keywords in AI search engines.

Options

DataForSEO requires both a location and a language. When omitted, location_code defaults to 2840 (United States) and language_code to en; use --location-code/--language-code to override.