AI Visibility Check

How many full audits the account has left, which of them expire and when, when the next free one is added, and the limits on quick audits. Read-only: it cannot buy or add audits.

check_audit_allowance

From a finished full audit: whether AI models (ChatGPT, Claude, Gemini and others) recognise the brand when asked directly, whether they name it unprompted when asked for a recommendation, per model, and which competitors they name instead.

get_ai_visibility

What to fix, in order, from a finished audit: each action with the problem it closes, the effort, and how much it adds to the score, plus the score after each fix.

get_action_plan

Progress of a running audit, or the summary of a finished one: score, band, headline, the top five actions and the link to the full report. Waits up to waitSeconds (default 20) for a running audit to finish before answering.

get_site_audit

The problems a finished audit found, most important first, 20 at a time. Filter by severity or by section (a section id or part of its title).

get_audit_issues

This account's recent audits, newest first, optionally for one domain.

list_site_audits

Every check in one part of a finished audit, with its result and evidence. Pass a group id — "found" (can AI and search engines reach the site), "understood" (do they understand it), "experience", "trust" or "technical" — or a section id or title such as "structured data", "AI crawlers" or "performance". An unknown name returns the list of sections.

get_audit_section

Start an audit of how well a website can be found, read and recommended by AI assistants and search engines. Returns at once with a scanId; poll get_site_audit for the result. depth "quick" checks the page itself in about a minute and doesn't use the account's allowance. depth "full" also crawls the site, measures speed, and asks AI models about the brand; it takes about 7 minutes and uses one of the account's monthly full audits, so confirm with the user first. If the site had a full audit in the last day, no new one starts: the answer points to that report, and fresh: true runs a new one when the user asks. If a full audit of the site is running right now, the answer points to that run instead of starting a second one.

start_site_audit