Check whether Claude recommends you when it searches the web.

PromptSignal measures Claude through its API with web search on, and shows whether it names you, who it names first and which pages it cites.

Illustration with sample data: Claude answers a buyer's question, “ What's the best project management software for a creative agency?”. PromptSignal reads the answer for a fictional brand, Tallyfield: mentioned at position 2 of 3.

Claude is measured through its API with web search.

The other six engines are captured from the interfaces buyers use. Claude is asked the same prompts outside the Claude app, and its answers are scored the same way.

Same metricsVisibility, average rank, first mention and sentiment for Claude, next to the other engines.

Kept in fullEvery Claude answer is kept with its citations, your position and the run history.

Illustration with sample data: seven AI engines answer a buyer's question, “ What's the best project management software for a creative agency?”, for a fictional brand, Tallyfield. PromptSignal reads each answer: Tallyfield is mentioned by 5 of 7 engines, at an average position of 1.8.

Claude searches when a question needs current information.

Anthropic says Claude searches when a question is about specific organizations or products that may have changed, and answers settled questions from what it already knows. When it searches, its answer cites the pages it used.

Citation gapsPages Claude cites that name your competitors but not you.

BriefsContent briefs for the prompts where Claude leaves you out, turned into articles in the editor.

Illustration with sample data: the Sources page lists the domains AI engines cite most for a fictional brand's prompts. Three of them name its competitors but not the brand; the first, g2.com, is opened to show the gap and a button to draft a brief.

Anthropic keeps its search crawler apart from its training crawler.

Claude-SearchBot indexes pages for Claude's search results, and Anthropic says blocking it may reduce your visibility there. ClaudeBot collects training data, Claude-User fetches pages when a person asks, and robots.txt allows or blocks each one separately.

Questions. Answered.

Is Claude's answer the same one a buyer sees in the Claude app?

Not always word for word. PromptSignal asks Claude through its API with web search, not through the Claude app, so phrasing can differ from what someone sees there. What you track is how Claude answers your prompts when it can search the web: who it names, in what order and which pages it cites.

Does Claude always cite its sources?

When it searches the web, yes: citations are always on for web search in Claude's API. When Claude answers from what it already knows, there are no pages to cite.

Can we compare Claude with ChatGPT?

Yes. The Models scorecard puts Claude beside ChatGPT and the other engines, with your visibility, average rank and first mention in each, and marks your strongest and weakest engine.

See what AI says about your brand.

Start a free trial and track your first prompts today, or book a demo on your own brand.