AI engine tracking
Not an estimate and not a model of what an engine might say. Real prompts, real answers, recorded.
The engines
ChatGPT, Grok, Perplexity, Gemini, Copilot and Claude. Answers are captured through managed browser automation rather than APIs, because the API answer and the consumer answer are not the same answer.
What comes back per answer
Whether your brand was mentioned, how many times, its position in the answer, a sentiment score, which competitors were named, and which sources the engine cited.
Competitor names are reconciled, not invented
Competitor mentions are matched against brands you have registered. If an extraction returns a name we do not recognise, it goes to a separate unknown brands list rather than being treated as a competitor you told us about.
Questions
How often does it run?
On a schedule you set per workspace, and on demand. Answer engines are not deterministic, so a single run is a sample rather than a measurement.
More in Connect and scale
- OAuth connection, real API reads
Search Console and GA4
Connect Search Console and Google Analytics 4 so ranking and traffic data sit next to the audit findings instead of in another tab.
Read more - Tenant id is the first argument, always
Client workspaces
Every client is a tenant. Isolation is enforced by the shape of the code, not by remembering to filter, and a record belonging to another client answers 404.
Read more - AES-256-GCM, per client key via HKDF
Credential vault
Client credentials are encrypted with AES-256-GCM under a per-client key. No endpoint returns a plaintext secret, including to you.
Read more
Run it against your own site.
No signup, no card. It reports what it could not check as well as what it found.