source
CUVETSMO · verified vet knowledge

Privacy stance

We don't want your data. Here's how that works in practice.

Most medical reference platforms make money from you (Plumb's/UpToDate paywalls) or about you (ad-supported sites tracking what drugs you look up). We do neither. This page is the explicit version of that — not legal boilerplate, an actual stance.

No analytics. No tracking pixels. No fingerprinting.

There is no Google Analytics, no Plausible, no Fathom, no Hotjar, no Mixpanel. No third-party JavaScript runs on these pages. Your browser tab's DevTools → Network panel will confirm: only assets from source.cuvetsmo.com+ (eventually) Vercel's edge.

We give up the ability to know what drugs are most-viewed. We accept that trade.

No cookies. No localStorage tracking. No IDs of any kind.

The service worker uses caches.open() for offline content (cached drug entries + UI shell). That cache lives only in your browser, never sent to us. No cookies are set. No localStorage / sessionStorage values are used for user identification.

Free public reads — no signup, no API key, no rate-limit identity

Every /api/* endpoint is open. CORS-enabled. No Bearer token required for read. Rate-limit headers identify your IP at the edge (Vercel CDN); we do not log or persist it server-side.

Phase 1+ may introduce optional API keys for institutional tier (bulk export, contribute back). Those will be opt-in and the public read tier stays anonymous.

Server logs: edge-level only, kept ≤ 7 days

Vercel keeps standard CDN access logs (URL, response code, timestamp, IP, user-agent) at the edge for ~7 days for operational reasons. We do not augment, export, or analyze these. They roll off automatically.

Your reading list never leaves your browser

Search queries, bookmark navigation, scroll depth, time-on-page — all stay local. The PWA caches content, but the cache index is not transmitted. If you save the page for offline (Ctrl+S / "add to home screen"), nothing is sent.


What we DO collect (transparency > pretending zero)

การ collect ที่หลีกเลี่ยงไม่ได้

  • Edge CDN access logs (Vercel, ~7 days, then auto-rolloff). Standard HTTP request metadata. We do not query or export these.
  • GitHub commit history. Every change to content/ is a public git commit by a named GitHub user. This is by design — the editorial audit trail is itself the trust mechanism. If you contribute, your GitHub identity is visible in the commit log forever.
  • Faculty signatures. When a faculty member signs an entry, their name + title + affiliation + date are committed to the entry JSON + the transparency log. This is also by design — accountability requires identity. Faculty consent to this when generating their key.

Never

เราจะไม่ทำ

  • Sell or share IP/user data with third parties
  • Add advertising of any kind
  • Track reader behavior across pages
  • Require an account or signup for public reading
  • Hold reader content (notes, bookmarks, queries) on our servers
  • Integrate third-party JavaScript SDKs for tracking

For context

เทียบกับเครื่องมือที่ใช้กันอยู่

PlatformAnalyticsCookiesAccount requiredCost
ChatGPT/Geminiyesyesyes$20-30/mo
Plumb'syesyesyesUSD 300/yr
UpToDateyesyesyesUSD 500+/yr
MIMS ThailandyesyesyesFree w/ login
DailyMed (FDA)public-sectorsomenoFree
WikipediaaggregatesomeoptionalFree
source.cuvetsmo.comnononoFree public

ตัวเลขนี้คือสิ่งที่เราเช็คในเดือน 2026-05 ผ่านการเปิด DevTools + อ่าน privacy policy ของแต่ละ platform.


The honest question

แล้วเราจ่ายค่า server ยังไง?

Phase 0: server cost ใกล้ศูนย์ — เว็บนี้คือ static content + edge-cached API responses. Vercel hosting tier ฟรีอยู่ในวง free quota.

Phase 1+: revenue ผ่าน institutional API tier (hospital chains, AI startups, government integrators ที่ต้องการ bulk export + SLA + dataset DOI). Revenue ส่วนนี้กระจายกลับสู่ภาควิชา ของ faculty ผู้ contribute. Public read ยังฟรีตลอดไป.

Iron Rule: เราจะไม่เก็บเงินจาก readers, ไม่เก็บข้อมูล readers, ไม่ขายอะไรกับใครเกี่ยวกับ readers. ดูรายละเอียดที่ /api ภายใต้ “Inverted economics”.