Documentation

Tools exposed by Website Crawler (SEO) and the limits that apply.

Connect

Server URL: https://crawler.tidybutton.com/mcp. Authentication is OAuth with dynamic client registration and an email sign-in code. Claude: Settings, Connectors, Add custom connector.

Tools

`start_site_audit`

Starts a crawl of a site from a start URL and returns an audit id. Crawls same-host pages, respecting robots.txt.

Inputs: url (public http or https URL), max_pages (optional, capped by the plan)
Limits: Counts one audit against the monthly allowance. Page cap per plan.

Audit https://example.com and tell me the biggest problems.

`get_site_audit`

Returns progress for an audit and advances the crawl by one batch. When finished, returns the score, counts, top fixes and the report URL.

Inputs: audit_id
Limits: Each call fetches a bounded batch of pages.

How is audit abc going?

`list_site_audits`

Lists recent audits for the account with status, score and report URL.

Inputs: none
Limits: Latest 20.

List my site audits.

`check_page`

Checks one page immediately: status, redirects, title, description, headings, canonical, noindex, images, structured data, Open Graph, weight, mixed content.

Inputs: url
Limits: One page, 8 second timeout.

Check https://example.com/pricing for SEO problems.

`account_status`

Shows the account email, plan, usage and limits.

Inputs: none
Limits: none

Show my account status.

`activate_license`

Binds a licence key to the account.

Inputs: license_key (string)
Limits: Key must belong to this product.

Activate licence key XXXX-XXXX.

Plan limits

Limit	Free	Windows app owner	Cloud Pro
audits	3	10	50
pages per audit	25	200	500

Network behaviour

Outbound requests use the user agent TidyButtonBot/1.0 (+https://crawler.tidybutton.com/bot), time out after 8 seconds and are blocked from private, loopback and link-local addresses.