Last Updated: August 26, 2026
1. Overview
Kids Scroll ("we," "our," or "us") is designed to be a safe, educational environment for toddlers and young children. We take children's privacy extremely seriously and have built this application to comply with the Children's Online Privacy Protection Act (COPPA) and the General Data Protection Regulation (GDPR-K).
2. No Information Collection
We do not collect any "Personally Identifiable Information" (PII) from children. This includes, but is not limited to:
- Names or Usernames
- Email addresses or Physical addresses
- Phone numbers
- Photos, videos, or audio recordings of the child
- Geolocation data
Our app does not require registration, user accounts, or any form of login.
3. No Advertisements
Kids Scroll is 100% ad-free. We do not include third-party advertisements, behavioral advertising, or any form of "in-app purchases." This ensures that children cannot accidentally click on ads or be exposed to commercial tracking.
4. Usage Data & Analytics
To improve our app, we use Google Analytics to collect non-personal, aggregated usage statistics. This data is used strictly for "support for internal operations" as permitted under COPPA. This includes:
- Which games are played most frequently
- How long users spend in the application
- Technical data such as device type and screen resolution
This data is completely anonymous and cannot be used to identify or contact any individual child. Furthermore, all ad personalization, tracking, and third-party data sharing features are completely disabled.
5. Stories: On-Device Generation & Local-Only Stats
Kids Scroll also includes an offline story feature. Stories are generated by a lightweight on-device micro-language model based on Microsoft Research's TinyStories architecture running via WebAssembly (wllama / llama.cpp). No prompt, story text, or listening activity is ever sent over the network to generate a story. After the first successful load, story time works with the device's network connection fully turned off.
The Parent Dashboard's story stats (how many stories were listened to, and for how long) are calculated and stored only in your device's local browser storage (IndexedDB). Like all data described in this policy, this never leaves your device and is automatically cleared if you clear your browser/app data.
If a story doesn't make sense, the 🚩 button in the story reader saves that story's text locally, viewable in the Parent Dashboard so you can copy and report it yourself if you'd like - it is never sent anywhere automatically.
6. External Links
To prevent children from leaving the app, we do not include links to social media or external websites within the games or stories. The only external links are located on this Privacy page and the "About" page, intended for parents' use.
7. Parental Rights & Data Deletion
As we do not collect any personal information or require user accounts, we do not store any personal data or account information on our servers. Therefore, there is no personal data to delete.
However, we respect your rights regarding any potential data. If you believe we have inadvertently collected any personal information, or if you have any concerns, you can request data deletion by contacting us at support@offlinekidsgames.com. We will verify and permanently delete any such information within 30 days.
8. Contact Information
If you have any questions about this policy or our privacy practices, please contact the developer at:
Email: support@offlinekidsgames.com
9. Development & Attribution
Kids Scroll is independently built and maintained by a single parent-developer, with AI coding assistants (Claude Code by Anthropic, Google Antigravity, and Gemini CLI) used as co-development tools throughout the codebase. These tools are used only to help write and review code during development - they play no role in the app itself and have no access to, or involvement in, how your child's data is handled at runtime.
Open Research & Component Attribution: On-device story generation utilizes the open research architecture and toddler-focused synthetic vocabulary principles from Microsoft Research's TinyStories ("TinyStories: How Small Can Language Models Be and Still Speak Coherent English?" by Ronen Eldan and Yuanzhi Li, 2023) and the open-source wllama / llama.cpp WebAssembly inference engine.