AI Visibility / GEO Monitor
Pricing
from $0.01 / 1,000 dataset items
Go to Apify Store
AI Visibility / GEO Monitor
Measure whether Claude mentions your brand in a live web-search-grounded answer, with sentiment and cited sources.
AI Visibility / GEO Monitor
Pricing
from $0.01 / 1,000 dataset items
Measure whether Claude mentions your brand in a live web-search-grounded answer, with sentiment and cited sources.
Brand names to look for in each generated answer, e.g. ["Acme", "Globex"]. Required, non-empty, max 20. Every brand is checked against the SAME generated answer for a given query/run (cheap: one classification call per brand), so adding brands does not multiply the expensive web-search+generation cost.
Search-style questions to ask the AI model, e.g. "best project management software for small teams". Required, non-empty, max 20. Each query costs one live web-search-grounded generation call, shared across every brand in 'brands'.
Which AI engine(s) to query. LAUNCH SUPPORTS ONLY "anthropic" (one Claude model with web search) -- this does NOT query ChatGPT, Gemini, or Perplexity, despite accepting a list for forward-compatibility with future engines. Any other value fails the run with a clear error.
[ "anthropic"]Which Claude model generates the answer. "claude-sonnet-5" (default) is the cost/quality balance this actor is priced around. "claude-opus-4-8" is higher quality and more expensive. "claude-haiku-4-5" is cheapest but does NOT support the dynamic-filtering web search tool this actor uses -- only usable with 'search': false (no live grounding).
How many times to independently re-ask each query (for variance measurement -- AI answers vary run to run, especially with live web search). Each run bills one shared 'answer-generation' event plus one 'brand-check' row event per brand. Must be between 1 and 5.
Whether the model may search the web for its answer (recommended: true -- this is the whole point of a GEO/AI-visibility check, since it measures what the model says RIGHT NOW, not just from its training data). Set false only to test cheaper/ungrounded answers or when using 'claude-haiku-4-5' (which doesn't support this actor's web search tool version).