Google Places API
Business listings and locations, ahead of the open-source Maps scraper.
The facts a reader needs
| What it unlocks | Nothing on the hosted report — it is used only by the local brand-md collection lane. |
| When it is called | Never in this deployment — GOOGLE_PLACES_API_KEY is not set, and the hosted lane has no path to it at all. |
| Cost basis | Billable per text-search query when configured; nothing has ever been billed on this account. |
| Environment | GOOGLE_PLACES_API_KEY |
| Bring your own key | Not on this deployment. |
The LOCAL brand-md collection lane only (src/collectors/provider-context.mjs). The hosted fourteen-leg pull never imports it.
Endpoints we call
No endpoint of theirs is called on this deployment today.
Brandfetch
A brand's logo when the site's own markup does not carry one.
Public web sources (no account, no key)
Nothing — these are endpoints anyone may call: Wikipedia and Wikidata, Google News RSS, Reddit search, Bluesky, the Apple and Google app stores, the iTunes podcast index, Trustpilot, the platforms' own ad-transparency centers, Google Autocomplete, and the platforms' oEmbed endpoints. yt-dlp, an open-source CLI, reads YouTube and TikTok the same way. The free-sources round (2026-08) added two more to this same no-key list: GDELT DOC 2.0 (news mention volume with a tone score) and the archive.org Wayback Machine CDX API (the site's nearest snapshot to a year ago, plus its oldest snapshot).