Is your brand visible in ChatGPT & AI search?

Do it by hand. Write 5 to 10 questions a real buyer would ask, such as "What are the best [category] tools for [use case]?". Ask each one in a new chat, so earlier messages don't steer the answer, and record whether your brand is named, its position in the list, and which websites are cited. Repeat each question two or three times, because ChatGPT can answer the same prompt differently. It costs nothing but is slow to repeat every week, which is the part the actor behind this page automates.

This page is free: it builds the query and shows a fixed example of the output, but it does not query ChatGPT from your browser. The LLM Visibility Monitor actor runs the real check. Its demo mode writes clearly labelled fake answers, calls no AI model and charges nothing, so you can see the output format first. Live runs are pay-per-event, and new Apify accounts get free platform credits to start.

Pick a fixed set of buyer-style prompts, run it on each AI engine, and count brand mentions. Share of voice = your mentions divided by (your mentions + the mentions of every competitor you track) across the set. Calculate it per engine, since each model answers differently, and record citations (linked sources) separately from mentions (your name in the text).

Run a fixed set of buyer-intent prompts through ChatGPT on a schedule and parse each response for your brand name, its position among competitors, any citations to your domain, and sentiment. Doing this manually is unreliable because answers vary per session, so most teams automate it. This builder generates the query and previews the output shape; the backing LLM Visibility Monitor actor runs the prompts live across ChatGPT and four other engines.

No. This is a query builder. You enter your brand name and market category, and it produces a ready-to-run configuration plus a fixed example of the output shape so you know exactly what fields you will get. The live querying of ChatGPT and the other engines happens when you run the backing actor on Apify, which calls the AI platforms' APIs. That live querying runs server-side in the actor, where the API keys live, not in your browser.

A mention means the AI named your brand in its answer text. A citation means the AI linked to your website as a source. They require different strategies and are tracked separately. The actor's output records brandMentioned (boolean) and brandPosition for mentions, plus ownedDomainCited (boolean) and citedUrls/citedDomains arrays for citations, so you can see whether you are being talked about, linked to, or both.

Five major AI engines: ChatGPT (OpenAI), Claude (Anthropic), Google Gemini (used as a proxy for Google's AI Overview signal; the actor queries the Gemini model, it does not scrape Google AI Overview SERPs), Perplexity, and Grok. You can select which platforms to query when you run the actor. Checking multiple engines matters because brand visibility differs sharply between them; a brand recommended in Perplexity may be absent from ChatGPT for the same prompt.

ChatGPT answers vary due to probabilistic token sampling, per-user personalization and memory, the model version in use, and live retrieval. In practice a brand's mention rate can swing widely from one week to the next without changing a single page, purely from sampling and retrieval drift; treat any specific percentage as illustrative of that volatility, not a fixed benchmark. That variability is why a single manual check is misleading and why recurring, multi-sample prompt runs are needed to read a real trend.

The builder on this page is free to use. The backing Apify actor has an optional demo mode (off by default) that writes labelled fake answers, calls no AI model and charges nothing. A normal run is live, and pricing is pay-per-event: about $0.15 per standard LLM sample and $0.35 per premium sample per the actor's pricing, and samples that return no usable evidence are not charged. New Apify accounts also get platform-level free credits to start.

Share of voice is your percentage of inclusion in AI answers across a defined prompt set, often called your market share of answers. The actor calculates it as your brand mentions divided by total brand plus competitor mentions. Unlike traditional SEO rankings, you are competing for inclusion in the AI's final response, not for a position in a list of blue links, so a higher share of voice means the AI recommends you more often than rivals.

Most GEO practitioners recommend monthly or quarterly audits, with higher-cadence monitoring for competitive categories. Because AI answers drift between sessions and model updates, run the same prompt set repeatedly and compare snapshots over time rather than relying on one check. A common pattern is to note when you publish content or earn press, then look for movement in mention rate and citations 4 to 8 weeks later.

Yes. When you run the actor you can supply a competitor list by brand name or URL. The output then tracks competitor mentions, position comparisons, and gaps where a competitor appears in the answer but your brand does not. This is what powers the share-of-voice calculation and the COMPETITOR_SHARE_OF_VOICE and PROMPT_GAPS summary reports, so you can see exactly which prompts your rivals own and you are missing.