How much does AI brand monitoring cost?

It depends on volume, because the underlying cost is paid API calls to the AI providers. A tool that reads a hundred prompts a week against two models is a fundamentally cheaper thing to run than one that reads a thousand a day against six, and pricing reflects that. The pricing page sets the same numbers out in full.

Because vendors are selling different query volumes under the same three words. Ask any vendor for prompts per period, providers queried, brands included and refresh frequency, and most of the price gap resolves itself. The rest is usually packaging: enterprise reporting, seats, multi-market coverage, agency workspaces. Our guide to choosing an AI visibility tool lists the eleven questions worth asking before you compare two numbers.

No. Free AI visibility checkers do exist and they answer a narrow question well, which is whether a model names you today. Our page on free checkers explains what a one-off scan can and cannot tell you, and why a single reading is not a trend.

Not on this plan. The limit is one brand and up to 3 named competitors per subscription, which keeps the prompt set tight and the score comparable week to week. Agencies running several client brands need one subscription per brand, and our agencies page explains how that works in practice.

We would rather you learn this here than in procurement. There is no independent security audit or certification, so no SOC 2 and no ISO 27001. There is no single sign-on: people sign in with email and password. There is no signed data processing agreement on request, no self-serve export of your stored answers and scores, and no choice of hosting region, because the app and its Postgres database run in a single AWS region.

Billing is monthly. You can delete the account yourself from Brand settings, which cancels a live subscription first and then removes the brand, its prompt set, every stored answer, the extractions, the scores and the digest history at once. If you cannot sign in, write to privacy@luminary.app from the address on the account. Records Stripe must keep for tax purposes stay with Stripe after deletion.

The variance is real, and it is the reason to measure on a schedule rather than by hand. Seenby runs the same frozen prompt set every week, so movement in the trend is not movement in the question. Every stored answer carries its model version, which is how a genuine shift in how you are described is separated from a provider changing the model underneath. The caveat travels with the score: week-over-week shifts may reflect a model update rather than a real competitive change, and that disclosure sits on every score display, in the digest, and on a dedicated methodology page.