Privacy

1. What we record, and why

Every search, suggestion and AI answer is written to our own database with these fields:

• Your search terms and AI questions — measuring result quality, finding queries that return nothing, deciding what to improve
• Timestamp — usage trends over time
• IP address — abuse and rate-limit investigation, and estimating distinct visitors
• A hash of the IP address — grouping activity for statistics without reading the address itself
• Country or region — understanding traffic sources and enforcing regional availability
• Browser user agent — device and compatibility breakdown
• Interface and search language — language usage breakdown
• Request path, status code, response time and size — performance monitoring and fault diagnosis
• Referring page — understanding how visitors arrive
• Result count, cache hit, and whether the upstream provider was called — service quality and upstream API cost tracking
• Which result you clicked, and its position — measuring whether the top results are the useful ones
• The full list of results shown for each search — reconstructing what a visit looked like when investigating a complaint or a quality problem
• Time engaged with the page and scroll depth — telling a five-second bounce apart from a five-minute read

2. Cookies and browser storage

• reflib_vid — a random identifier for your browser, kept two years, used to tell repeat visits apart. No name, e-mail or account — there are no accounts here
• reflib_sid — a random identifier for the current visit, expiring after 30 minutes idle, used to group one sitting together
• reflib-theme / reflib-lang — your own theme and language choices; never sent anywhere for analysis
• reflib-history (browser storage, not a cookie) — the last 50 queries you typed, kept only in this browser to offer back in the search box. Never uploaded; clear it from the dropdown under the search box

3. How we meet the six Data Protection Principles

• DPP1 — Purpose and manner of collection — collected only to run the service, prevent abuse and improve the product; not excessive; this page is the notice given before collection
• DPP2 — Accuracy and duration of retention — recorded automatically and not edited by hand; retention is set out in section 4
• DPP3 — Use of personal data — not used for any new purpose, not used for direct marketing, and never sold or rented to anyone
• DPP4 — Security — HTTPS enforced site-wide; the admin dashboard sits behind Cloudflare Access with e-mail verification; the database is not reachable from the internet and only named operators can read it
• DPP5 — Openness — this page states our policies and practices, and is linked from the footer of every page
• DPP6 — Access to personal data — you may request access to, and correction of, the data we hold about you — see section 6

4. How long we keep it, and where

• Detail records are currently kept indefinitely; we do not delete them on a schedule. Retention limits are being introduced and this page will be updated to match when they are.
• A separate aggregate layer (daily counts by region and language) holds no IP address, visitor identifier or search term, cannot be traced to any individual, and is kept long-term to support trend analysis.
• Data is stored on our own server in Hong Kong.
• The database is backed up daily and backups are kept for 30 days.

5. Who can see it, and who it goes to

• Only the people who operate this service can read the raw records, for operations and analysis.
• We do not sell, rent or supply personal data to third parties for marketing.
• We will disclose data where required by law to Hong Kong law enforcement or a court, or where necessary to protect the legitimate interests of the service, its users, or the public.

6. Your rights (DPP6)

You may ask whether we hold personal data about you, request a copy, and require correction of anything inaccurate. Write to [email protected].

Because the site has no accounts, we hold no name or e-mail to identify you by. You will need to give us enough to locate the records — for example your visitor identifier (the value of reflib_vid), the approximate time of the searches, and the IP address you used. If we cannot match a request to specific records, we cannot comply with it.

As the Ordinance requires, we will respond within 40 days; if we cannot comply, we will say why within the same period.

7. Upstream services and cross-border transfer

The service depends on the following third parties, and your queries pass through their systems:

• Brave Search API — performs the actual search. Your search terms are sent to its servers and handled under its own privacy policy. We send it no identifier beyond what the connection itself reveals
• Cloudflare — content delivery, DDoS protection and regional rules. Every request passes through its network, which handles your IP address and region

• Servers of those services may sit outside Hong Kong, so your queries involve a cross-border transfer. By using this site you acknowledge and consent to that transfer.

8. What we do not do

• No accounts; we never ask for your name, e-mail or phone number.
• No third-party advertising or analytics code (no Google Analytics, no Facebook Pixel).
• No selling of data and no direct marketing.

9. Complaints

If you have a question or complaint about how we handle personal data, write to [email protected].

You also have the right to complain directly to the Office of the Privacy Commissioner for Personal Data (PCPD) in Hong Kong.