Skip to content

Privacy Policy

For the OK JLPT (OK真题) iPhone, iPad and Android App, this website and OK JLPT in ChatGPT. Updated October 28, 2026.

1. Operator and contact

The service is operated by OK JLPT, based in Hong Kong. Contact support@okjlpt.com for privacy, data rights and support. This policy covers the JLPT product; other OK Nihongo products have separate policies.

2. Data we process

Account data: sign-in email, shared identity ID, sign-in method and account status. Profile data: nickname, bio, an avatar if uploaded, study identity, target level, exam date and time zone. Learning data: answers, drafts, time spent, score estimates, mistakes and review or practice progress.

Interaction data: AI questions and replies, learner notes, likes, reports, blocks and acceptance of community rules. Purchase and activity data: Apple, Google Play and Stripe transaction identifiers, products, purchase and refund status, access rights, AI balances, invitations and eligibility records.

Security and diagnostic data: request times, IP addresses or risk summaries, installation identifiers or hashes, App version, device platform, errors and request IDs. Notifications also use a push token and your preferences.

Usage analytics: App lifecycle, page visits, paper and practice starts and completion, purchase actions and AI feature use. Events may include level, paper or question IDs, question counts, correct-answer counts and time spent. After sign-in, events can be linked to your account ID.

3. Purposes and choices

We use this data to manage accounts, sync learning, provide explanations and AI responses, run and moderate learner notes, verify and restore purchases, assess invitations, send reminders you permit, resolve errors and prevent abuse. Optional features do not require your active input when you do not use them. We do not sell personal data for advertising.We also use analytics events to understand feature use, check study and purchase flows and improve the product.

4. AI and moderation

When you ask AI or generate a practice question, your text, the relevant question and explanation, and necessary context are sent to Google Gemini. Conversations are saved to your JLPT account for later review. Do not include passwords, verification codes, identity documents, health information or other people’s private data.

Notes may be checked by content rules and AI-assisted moderation; note content may be sent to Google Gemini for that purpose. The provider’s applicable terms and privacy practices also govern its processing. AI can make mistakes. Using the product does not authorize undisclosed data uses.

5. Public content

Published notes, nicknames, avatars and public profile fields can be viewed by other learners. Sign-in email, private drafts, answers and AI conversations are not published as learner notes. Avoid sharing sensitive data in notes or profiles. You can delete your notes, report content and block users. The operator may access relevant records where necessary to resolve reports, security issues or support requests.

6. Providers and international processing

Supabase manages shared sign-in identities; Apple or Google handle their selected sign-in methods; Apple and Google Play process in-app purchases and Stripe processes website payments; Google Gemini provides AI generation and some moderation; Cloudflare stores or delivers media and avatars; AWS provides server hosting; Better Stack receives privacy-filtered diagnostic logs, which may include account IDs, email HMAC fingerprints, App versions and request IDs, without plain-text email addresses or request bodies.

Necessary data may be processed outside Japan depending on provider infrastructure. We restrict access and use encrypted transport. Providers independently process data they collect under their own policies.

PostHog receives usage analytics for key actions in the App and on the website (such as submitting a paper or starting a purchase) with associated user identifiers through its US endpoint; the website sets no cookies for this.

Ahrefs counts website visits and where they come from (such as search engines); it sets no cookies and is not linked to your account.

The website notes in your browser which site or link brought you here (the site's name, the campaign and the page you landed on) and saves it to your account after you sign in, to learn which channels bring learners and purchases; the full referring address is not kept.

When you use OK JLPT in ChatGPT, we receive only what a question needs (level, section, the answer you chose, language) and an anonymous ID from ChatGPT, used to count usage; no sign-in, and no chat content, name or email address.

7. Retention and deletion

While your JLPT account exists, we retain records needed for learning sync, conversation history, access and community features. JLPT deletion clears its learning records, drafts, mistakes, profile, notes, likes, AI conversations and balance through the account deletion process.

The shared OK Nihongo sign-in identity remains so other products are unaffected. Necessary purchase records (Apple, Google Play, Stripe) are unlinked from the JLPT account and retained for three years under the product retention policy, then deleted, unless law requires longer retention. Invitation eligibility, duplicate-reward prevention and deletion records remain while needed to prevent abuse. Report and security records may remain until resolution and necessary review are complete. Backups and provider records may not disappear immediately.

8. Your rights and controls

Use My page / Delete account in the App to review the scope and confirm JLPT account deletion. It cannot be undone, does not automatically refund purchases and does not delete your Apple or Google account. Remaining AI credits are cleared; restoring a pass depends on valid purchase records.

Email support@okjlpt.com to request access, correction, a copy, processing restrictions or shared-identity deletion. We verify ownership and explain effects on other products or required retained records. Notification permission can be withdrawn in system settings. Avatar upload uses a photo you select; listening playback does not require recording your voice.

9. Children, security and changes

The service is for Japanese learners and is not specifically directed at young children. Minors should obtain guardian consent where applicable law requires it. Contact us about improperly collected children’s data. We use reasonable security measures but cannot guarantee absolute security. Material changes to data uses or providers will be reflected here, with notice and consent where required by law or platform rules.