s-island

s-island is a safe island for your child, amid a sea of harmful content. It gives parents control over which sites are allowed or blocked, shows risk summaries (e.g. harmful content, personal info leak, predators), and can run optional AI checks on pages, while keeping messages and personal data private.

Parents see safety classifications and risk reports, but never the child’s chat content or private communications. Many existing tools force a choice between full surveillance and no real protection. s‑island is designed to avoid both extremes as detection and blocking happen in a private way. Only risk signals and domain level history are shared with parents while sensitive data stays local or is never collected at all. The system gives parents peace of mind: whitelists, blacklists, age‑appropriate rules, and clear risk labels, without crossing into surveillance.

The product consists of a web app for parents, a Chrome extension that runs on the child’s device, and a backend that stores devices, lists, and visit records. Parents sign up and log in on the website, create one or more devices (e.g. kids’ laptop), and receive an API key per device. They install the extension on the child’s browser and paste that key in the extension options. Parents can choose whether put their own prompt for AI agent in the extension or to let the child decide on the prompt themselves. Site visits are reported to the backend so the parent can see a visited list and risk flags in the portal, all without the parent ever seeing the child’s private information.

Built With

• amazon-web-services
• django
• docker
• exa
• fastapi
• featherlessai
• hive
• javascript
• manifest
• minimax
• nextjs
• postgresql
• python
• react
• sapling
• tailwind
• typescript