How to build an AI visibility dashboard on top of an API
Build a project per client, store prompt sets as reusable templates, schedule runs at a cadence matched to each client's budget, persist the full distribution rather than a point estimate, and trigger alerts only when a change exceeds the confidence interval. Keep raw answers so any number can be evidenced on demand.
Nostimates ships its own dashboard, so most teams never have to build one. But agencies and platforms often want their own — their brand, their data model, their client logins, wired into tools they already run. If that's you, here's a practical shape for building it on top of our API.
Data model
- project → client. Keys are project-scoped so one client's data can never leak into another's view.
- prompt_set → a reusable template. Onboarding client #41 should take minutes, not an afternoon.
- run → prompt × surface × country × timestamp, storing the full distribution, not a single number.
- sample → the raw answer text, retained for evidence.
Cadence is a cost dial
Frequency is a 7x multiplier between weekly and daily. So is surface count, and so is n. Give your account managers a per-client dial and let them spend more on the clients who pay for it.
Alert off the interval, not the point
Firing an alert every time a rate moves produces noise your team learns to ignore. Fire only when the new estimate falls outside the previous interval, and the alert means something every time it arrives.
White-label by default
See this for your brand
Nostimates shows you how your brand shows up across twelve AI engines and search — in a dashboard, or as data in your own tools. Tell us what you want to measure.