SafeCircle is building privacy-first AI for digital family protection — helping schools and families detect online risks in real time without invasive surveillance. Our app monitors conversations across platforms like WhatsApp, Instagram, and Telegram, sends instant alerts when something dangerous is detected, and is designed to protect children while preserving trust, transparency, and privacy.

We’re focused on a B2B-first model for educational institutions, because schools need practical, compliant tools to protect students online. That means multi-year contracts, easy deployment, GDPR-aligned data protection, and a system that fits into real school workflows instead of replacing them.

What makes SafeCircle different is that it doesn’t try to expose everything — it only surfaces the threats that matter, helping parents and schools act early without turning safety into surveillance. The product includes real-time alerts, a dashboard with high-level insights, and a family hub that encourages safer conversations at home.

At the core of this is Horizon, SafeCircle’s child safety risk detection model. Horizon is an on-device AI model built for conversation analysis, with a focus on detecting grooming, bullying, and threats in real time while keeping data local and private; the public model release describes Horizon Full GGUF variants and reports a Macro F1 of 0.8218 on a 160,000-example benchmark.

SafeCircle also has support and momentum from Cloudflare for Startups, NVIDIA Inception, EduCaixa / Fundación “la Caixa,” and CEEIM, plus a wider strategy around educational partnerships, EdTech distributors, school districts, and government education departments.

Long term, the goal is clear: become the trusted digital safety layer for schools first, then families, and help protect children online without compromising privacy.

🔍 Curated by @faizantariq26
✍️ Original thesis by @SafeCircleAI
🔗 https://x.com/SafeCircleAI/status/2081412525183164475