Is your brand in the AI answer?

Yes. Pay once and the app is yours, including free updates across version 1 — no subscription, no seats. The launch price is $29 for the first 100 customers; after that it's $79. The only ongoing cost is your own AI API usage — typically a few cents per daily run.

Click “Try free for 7 days”, verify your email, and the download starts — no payment details asked. The trial clock starts the first time you launch the app, and every install works for 7 days with no licence. After that it locks until a licence is added.

Yes — one OpenRouter key (openrouter.ai) runs all five engines with live web search. You pay OpenRouter directly at cost. The app shows a cost preview before every run and enforces a per-run budget cap (default $5, and daily runs usually cost far less), so spending can never run away.

You define the questions your buyers ask. The app sends each one to each engine with web search enabled, then checks the answer and its cited sources for your brand names and domains. Share of AI Answer is the percentage of question-engine pairs where you appear — cited results divided by results that completed without an error, so a provider outage never looks like a visibility drop — tracked over time, per engine and overall.

Fix a set of 10–15 questions your buyers actually ask, send them to ChatGPT every day with web search enabled, and check both the answer text and its cited sources for your brand. Asking once in the chat window tells you about one sample, not about your visibility — answers vary run to run. That daily loop is what this app automates, across five engines rather than one.

Because an ungrounded model answers from stale training data and fabricates citations that look completely real — right-shaped URLs for pages that never existed. Those fakes are indistinguishable from genuine citations once they are in a table, so grounding is not a quality setting here: it is the difference between measuring the answer surface and measuring the model's memory. Citations are read from each provider's structured grounding metadata, never scraped out of the answer text.

For the measurement, yes; for the platform, no. Those are cloud subscriptions where the vendor runs your queries and stores the results, with team seats, compliance artefacts and crawler analytics this app does not have. This does the same measurement loop — grounded prompts, citation checking, per-engine scoring, competitor gaps — as a desktop app for $29 once, on your own API key, with the data on your machine. The comparison pages state plainly where each platform wins.

Both. Apple Silicon Macs (M1 and later) on a recent macOS — the build is signed and notarized by Apple — and Windows 10/11 (64-bit). The Windows build isn’t code-signed yet, so it shows a "Windows protected your PC" warning on first run — click More info → Run anyway.

Yes. Your licence is your email, not a platform — it works on either macOS or Windows, one computer at a time. To move it between computers (including Mac ↔ Windows), deactivate in the app on the old one, then activate on the new one.

Nowhere. Prompts, answers, scores and API keys stay in a local database on your computer (keys encrypted with your OS keychain). The app talks only to the AI providers you configure — there is no server of ours in the middle.

Almost nothing. The app has no telemetry and sends us nothing — everything it does happens on your desktop, in your storage, under your control (backups included). On our side we keep only what a licence requires: your purchase email and an anonymised device fingerprint, with payments handled by Stripe. If you start a free trial from this site, we also keep the email you verified for the download. This website uses cookieless aggregate analytics, plus Google Analytics if you accept the cookie banner — neither can see anything the app does on your machine. Full details in the Privacy Policy and Terms below.