Search is being replaced by answers, and almost nobody can see what the answer says about them. MentionMachine asks all five engines the buying questions a customer would really type, and records whether you are absent, named, recommended, or cited as a source. The reply is stored verbatim, with the model version and the date it was fetched.
The hard part is honesty. An API surface is not the consumer app and can answer differently, so every result is labelled with the surface it came from. Nobody controls what an engine says, so the product never promises a ranking. It promises that the change was made correctly, verified, and re-measured. Most of the design work went into refusing to overclaim: the llms.txt check even says outright that it is newer and less established than the rest, and to treat it as cheap insurance.
A free scan runs the technical checks with no signup. The paid licence does the engine work, shows who gets recommended instead of you and which sources they cite, and generates a Fix Pack: schema, an llms.txt, a robots patch, FAQ answers and page briefs, built from the business's own prices and service area rather than invented. Every licence includes a verification re-scan, so the before and after sit side by side.
Priced as a $99 one-time lifetime licence rather than another subscription, metered in credits because every answer fetched costs real money: one credit is one question, on one engine, for one location. Cached answers are free, technical re-runs are free, and a cell that errors is refunded automatically.