Data Sources & Methodology

Business identity, ratings & status

When you add a competitor, tevralo looks it up through the Google Places API — the same official, licensed source Google Maps itself runs on. From it we get a business's name, address, location, price tier, star rating, review count, and open/closed status. We never scrape Google Maps or any other site to get this data; it comes through Google's own API under Google's terms.

We only ever read aggregate rating and review count — never individual review text — and Google's terms require us to re-check this data fresh on each scheduled check rather than archive it indefinitely, which is exactly how our monitoring already works (see below). Where this data is shown, you'll see a small “Rating & status data by Google” attribution, and your Market Map shows competitor locations on an actual Google Map.

Menu items & pricing

There is no legitimate, licensed API that provides itemized menu or pricing data for arbitrary independent restaurants — Google, Yelp, and every other major platform simply don't expose it. Menu and price tracking on tevralo comes from two places instead, and every item you see is labeled with which one produced it:

Owner-entered — from a competitor's profile page, you log a price you actually observed, or paste/photograph a whole menu and let Gemini structure it into items for you to review before saving. This is always available and always the most reliable option, since it's something you personally confirmed.

Automatically collected — for a competitor with a website on file, tevralo will attempt to fetch their menu directly from their own public site (checking their robots.txt first and only reading pages they've made publicly available — never a delivery app, never anything behind a login), and use Gemini to structure whatever text it finds into items, the same way it structures a pasted menu. This is best-effort: some sites can't be read this way at all, and AI extraction from a real web page is never perfect. Anywhere automatically collected prices are shown, you'll see a note that they may not always be current or 100% accurate — treat them as a helpful starting point, not a guarantee.

How monitoring works, mechanically

On a schedule (daily or weekly, depending on your plan — see Pricing), tevralo re-checks each monitored competitor's rating, review count, open/closed status, and (where applicable) menu against its sources, compares it to what we last recorded, and logs anything that changed as a market event. Owner-entered menu/price events are logged the moment you record an observation. Nothing is inferred or guessed; every event traces back to a specific observed change, and every data point records which source produced it (Google Places, a business's own website, or you) so it stays auditable from your Competitors page.

AI-generated insights

Market Insights and weekly report summaries are generated by Google's Gemini model, grounded only in the structured data described above — nothing else is sent to it, and it never sees your account credentials or billing information. It's instructed to say explicitly when there isn't enough data for a reliable insight rather than fill the gap with a plausible-sounding guess, and the product always shows whether a given insight came from the AI or a deterministic, data-grounded fallback (used automatically if the AI call fails or isn't configured).