Privacy policy.

This Privacy Policy describes how the Oreumics service (the “Service”) processes personal data. The Service has two surfaces:

• The desktop application you install on your computer (Windows; macOS / Linux planned).
• The web service at this site — used for account signup, Beta Access tickets, license validation, documentation, and support.

Different data flows apply to each surface. The desktop app is local-first and processes your scientific data on-device; the web service is the only component that handles personal data on our servers.

When you create an account, sign in, or redeem a Beta Access ticket, we collect the following categories of personal data:

Account profile

• Email address (used as your login identifier)
• Display name
• Institution / organization, department, and role
• Research fields and intended use case
• Country

The institution / role / research-field fields are used to understand which scientific communities use Oreumics and to prioritize features. They are not sold, shared with advertisers, or used for marketing profiling.

Authentication and session

• Hashed password and Supabase-issued JWT tokens (via @supabase/ssr)
• A session_id and device_id used to enforce the one-device-per-account rule
• Sign-in timestamps and IP addresses (for security audit only)

Subscription & billing

During the beta there are no paid plans and we collect no payment data — access is granted through free Beta Access tickets (next heading). When paid plans launch, payments will be processed by Paddle.com Market Limited (the Merchant of Record). Paddle is the data controller for payment information such as your card number, billing address, and tax identifiers. We never see your full card number; we only receive a Paddle customer ID, your active price ID, and subscription status to provision your plan. Refer to Paddle's privacy policy for how they handle that data.

Tickets (beta access codes)

If you redeem a beta-access ticket, we record which ticket you redeemed and its expiry date in your profile.

Email delivery

Transactional email (ticket-expiry warnings, password reset, etc.) is sent via Resend. We do not send marketing email by default. If we ever introduce a newsletter you will be asked to opt in explicitly.

The desktop app includes optional AI features — the AI Copilot chat, LLM-based Auto Annotation, and the report narrative writer. They stay off until you enter your own API key for the provider you choose (OpenAI, Anthropic, or Google Gemini) in the app's API Key Settings. The app calls that provider directly from your machine — Oreumics does not proxy, store, or see these requests. When you use these features, the app sends the following to your chosen provider:

• Your chat messages (the prompt text you type)
• For Auto Annotation: the per-cluster marker gene lists being annotated
• A summary of the current workflow — node graph structure, node parameter values, and dataset shape (e.g. cell count, gene count, list of clusters)
• Recent log lines and high-level analysis status (e.g. “UMAP completed with 15 PCs”)

The app does not send raw expression matrices, per-cell observations, sequencing reads, individual feature values, or your input files. Only the structural / metadata summary above leaves your machine.

If you do not want any data leaving your machine, simply do not enter an API key — the rest of the analysis pipeline functions identically without one. You can also clear chat history from within the app at any time.

Because the key is yours, the data relationship for these calls is between you and the provider, under its own API terms — for example OpenAI's privacy policy. As of writing, the major providers do not use API inputs to train their models by default.

The web service uses cookies only for what is strictly necessary for the site to function:

• Supabase auth cookies (sb-<project-ref>-auth-token) — keep you signed in.
• Preference cookies for things like the “remember my email” checkbox on the sign-in form.

We do not run third-party advertising trackers, behavioural-analytics SDKs (Google Analytics, Hotjar, Segment, etc.), or social-media pixels by default. If we add a privacy-respecting analytics tool in the future, this section will be updated and any meaningful change will be highlighted on the dashboard.

• Contract performance — we cannot operate an account or a Beta Access ticket without your email and ticket record.
• Legitimate interest — security audit (sign-in IP, device ID), fraud prevention, and improving the product (aggregate, non-personal usage metrics from the web only).
• Consent — for the AI features, which are optional and only active once you enter your own API key.
• Legal obligation — once paid plans launch, invoicing and tax records required by law (handled by Paddle as Merchant of Record).

• While your account is active: profile and subscription data are retained for as long as the account exists.
• When you delete your account: profile data is deleted within 30 days from our primary database. Backups are rotated out within an additional 30 days.
• Billing records (held by Paddle) may be retained longer where law requires (typically 5–10 years for tax purposes).
• AI Chat logs: chat history is stored locally on your machine. Server-side, your chosen LLM provider's API retention applies (for example, OpenAI states up to 30 days for abuse monitoring).

Our infrastructure providers may process data outside your country of residence:

• Supabase (account database) — the project region is configured at signup; data resides in that region.
• Vercel (web hosting) — global edge network.
• Paddle (billing) — processes payments via its global processor network; see Paddle's privacy policy for details.
• Resend (email) — US-based email infrastructure.
• Your chosen LLM provider (OpenAI, Anthropic, or Google — only when you use the AI features with your own key) — US-based inference under your own account with that provider.

Where personal data crosses borders, we rely on Standard Contractual Clauses or equivalent safeguards offered by each provider.

If you are in the European Economic Area, the United Kingdom, Switzerland, South Korea (PIPA), California, or another jurisdiction with comparable data-protection law, you have the right to:

• Access the personal data we hold about you
• Correct inaccurate data (rectification)
• Request deletion (the right to be forgotten)
• Receive your data in a portable format
• Object to or restrict certain processing
• Stop the AI features at any time by removing your API key
• Lodge a complaint with your local supervisory authority

To exercise any of these, email us at imza.kim@jejunu.ac.kr. We respond within 30 days.

Oreumics is a professional research tool intended for use by researchers, graduate students, and other adult professionals. The Service is not directed to individuals under 18 and we do not knowingly collect personal data from anyone under 18. If you believe a minor has created an account, please contact us so we can remove it.

We use TLS in transit, hashed passwords (handled by Supabase Auth), Postgres row-level security on every personal data table, and the principle of least privilege on service credentials. We rotate access keys regularly and restrict service-role access to server-side routes only. No system is perfect; we will notify affected users of material breaches in line with applicable law.

When we update this policy, the “Last updated” date at the top will change. For material changes (e.g. a new category of processing, a new sub-processor that affects most users) we will additionally surface a notice on the dashboard or by email.

Privacy questions and rights requests: imza.kim@jejunu.ac.kr

For general support, use the contact form or email gptapikorea@gmail.com; see also the FAQ.