Privacy · updated 7 October 2026

Privacy & learning

Pre-launch information: the verified seller/data-controller identity, postal address and support contact must be completed before registration, shared contributions or live sales open.

Your account

nena is the new name for nika. Existing account identifiers, subscription references and balances are preserved. Supabase manages your account email and authentication records. Our account service stores subscription and payment references, credit grants and AI usage metadata such as model, token counts, cost and request status. These records operate your account and account for credits; they are not a shared training dataset. Stripe handles payment-card details. nena does not store your card number.

Sources and AI generation

When you choose to send a source for generation, the content needed for the task is processed by nena’s service and its AI provider. Review the source before sending it. Do not submit patient data, confidential records or personal information about other people. Generated answers can be wrong and must be checked against the source.

Our OpenAI requests use store:false. This is not a promise of zero retention: provider abuse-monitoring logs may still be retained under its policies. OpenAI API data is not used for training its models by default. See OpenAI’s API data controls.

Desktop and cloud workspaces

The desktop app keeps card history, edits, source references, review feedback and personal-model files in an account-specific workspace on your computer. A web workspace saves the sources, cards and completed actions you explicitly submit to its processing service so that work can be resumed. Opening the web app does not automatically upload desktop history, Anki decks or local media. Importing desktop data or selected deck examples requires a separate action.

The web workspace is a staged beta. Its processing host and availability are configured for the particular deployment. The account portal uses Cloudflare and Supabase; a full cloud workspace requires a configured processing service. Cloud hosting, retention and deletion operations must be verified before a deployment is opened to customers.

Your personal learning

Your stated goals and study preferences belong to your account. You can inspect and change them. Card acceptance is not evidence of topic mastery, and deletion does not by itself show that a topic is unimportant.

With personal learning enabled, completed edits, additions, acceptances, mask corrections and deletion reasons are recorded as evidence. Eligible examples train the appropriate personal classifiers or neural-network components in the background. A candidate model only becomes active after its validation requirements pass. Frozen encoders and the hosted generation model are not fine-tuned after each action. A saved example is not necessarily eligible for every model.

Cloud personal learning requires its own permission in the workspace. It is separate from optional shared contributions. Relevant reviewed examples and saved preferences may be included in future generation requests when authorised. Turning off personal learning prevents further use of that evidence for personal adaptation; the service must recheck consent before activating a model.

Anki and question-bank access

The browser companion requires pairing with your account and specific deck/site permissions. It connects to Anki on your own computer. Selected-deck examples and queued question captures are sent to the workspace only when you choose to import them. Switching browser tabs does not end an active session, but observation may be unavailable while a page or the browser is closed.

Optional shared training

Shared contributions start off. If you opt in, you can review and submit individual card examples for future shared-model evaluation and training. Each contribution requires confirmation that it contains no personal or patient information and that you have permission to share it. Do not submit copied question-bank material, raw PDFs or screenshots as shared contributions.

Submitted examples are stored with a pseudonymous account identifier, timestamp, consent receipt and policy version. Pseudonymous does not mean anonymous. Other users cannot browse these examples. Shared-model training is not automated.

Changing your mind

Account → Withdraw & delete shared examples turns sharing off and deletes your stored shared examples from the active contribution database. Future datasets must exclude withdrawn examples. This does not automatically undo influence on a shared model already trained. Shared-model releases require reviewed consent and deletion/retraining procedures.

You can pause personal learning independently of shared contributions. Imported evidence, retained sources and model files must be included in a workspace data request; a general account export may not contain files held by a separately configured workspace service.

Access and account deletion

In Your data and privacy, download the account service’s JSONL export or submit a verified account-erasure request. We record a reference and a one-month response target. An erasure request immediately withdraws shared contributions, but it does not itself cancel billing or erase local desktop files. Support reviews outstanding billing, required retention and provider-held copies before closure. Records that must legally be retained are restricted and excluded from training.

Service providers

Supabase provides authentication and database services, Cloudflare serves the account portal/API and protects account forms, Stripe processes payments, and OpenAI provides AI inference. A workspace processing service may be configured separately. Provider locations and agreements determine where processing occurs; no UK-only or Europe-only processing promise is made here. Contact the verified controller above for access, correction or deployment-specific privacy questions.