Privacy Policy
The shortest, most aggressive privacy policy in academic software. No legal traps, no tracking telemetry, and zero manuscript ingestion.
The Engine
Your document is read inside your browser. The manuscript body — your unpublished research, drafts, data and prose — is discarded there and is never transmitted. What reaches our engine is the reference list itself: citations to published works, the same identifiers your browser already sends to Crossref. It is checked in memory and handed back, and the references themselves are never logged: your account keeps only a count of the checks you have run and the time of the last one. It is kept only if you choose to save the audit to your account, as described under Saved audits below.
Zero Tracking
StrictCite does not run Google Analytics, Plausible, PostHog, advertising pixels or similar tracking software. We do not fingerprint your browser, build behavioral profiles, sell usage data or set analytics cookies.
Cloudflare processes ordinary request metadata to deliver and protect the site. We also use short-lived security counters to prevent abuse of account features. This information is used for security and availability — not marketing, product analytics or user profiling.
The Payments
All billing and card transactions are handled securely off-site by Stripe. StrictCite never sees, transmits, or stores your credit card details or banking credentials.
Data Isolation Breakdown
What reaches our infrastructure versus what remains permanently offline.
- Manuscript text & drafts: Never submitted or uploaded.
- Unpublished lab findings & data: Zero exposure to any server.
- Bibliography segmentation: Regex-first, entirely client-side — GROBID's real trained model, run through Wapiti (GROBID's own engine, ported here to WebAssembly), steps in only on tangled cases, not a generative AI.
- Public DOIs & titles: sent from your own browser, on your own IP, straight to Crossref, DataCite, OpenAlex, DBLP, doi.org and OpenLibrary. They never pass through us.
- Registry responses: Cached ephemerally to avoid rate-limiting.
- Your reference list: sent to our engine, checked against the registries, and returned. On the free tier nothing survives the request.
- Saved audits (Pro and Lab): a bibliography is a list of other people’s published work — DOIs, titles and journals that are already public records. Cloud History keeps the finished report, and the corrections you approve, so you can reopen them on any device, and indexes those DOIs so we can warn you if a paper you cited is retracted next year. Neither feature can exist without remembering them; that is what the plan buys.
- Never your writing: the manuscript, the drafts, the unpublished data and results. Read in your browser, never sent — on every plan, including the paid ones.
- Yours to erase: Clear All in the workspace deletes every saved audit from our servers, and the retraction index goes with it. Nothing is kept back for our own records.
Why Commercial LLMs Cannot Scrape Your Research
Unlike cloud-based 'AI citation assistants' that ingest your full manuscript into OpenAI or Anthropic servers, StrictCite never receives your writing. Because your unpublished hypotheses, equations, and prose never leave your device, AI bots and web scrapers are physically unable to memorize or harvest your work.