ChatGPT search visibility: is your brand in the answer?
Give it the questions your customers ask and the brands or domains you care about. It asks ChatGPT's web search through the official OpenAI API, using your own OpenAI key, and returns one row per question: the answer, the URLs it cited, and for each brand whether it was mentioned, whether it was cited, and where it first appeared.
What it does
- Per prompt: answer text, cited URLs with titles, model name, timestamp, and the number of web searches the model made.
- Per brand or domain: mentioned (whole-word match in the answer), cited (a cited URL on that domain), mention count, character position of the first mention, and the order your tracked brands appear in. Add competitors to compare.
- Your key stays yours: it goes only to api.openai.com, is never logged or put in the results, and the run stops without a call if it is missing. OpenAI bills you directly for the searches and tokens.
- A cap on prompts per run (default 50, at most 200), a status on every prompt, and retries on rate limits. One failed prompt does not fail the run.
- Runs in 256 MB. Callable by AI assistants through Apify's MCP server.
Good to know
This uses the OpenAI API with its web search tool, not the consumer ChatGPT app. Answers and citations can differ from what people see in the app, so treat the results as a repeatable proxy, not a mirror. Matching is plain text: a misspelling is missed, and a mention can be a warning rather than a recommendation. You need an OpenAI API key that can use web search; there is no free tier here for the model calls.
Price
No tool fee yet: you pay Apify's platform usage for your run (a run with a rejected key cost under US$0.001 in our test) plus OpenAI's own charges. We have not run it against a live OpenAI key, so the first real runs may turn up issues; please report them.
A bug or a feature you need? Open an issue on the Store page.