SafeCircle

As a recap, great progress on Horizon for Android — now benchmarked across 60 devices. We’re continuing to improve performance and reliability across the Android ecosystem. Read more here: safecircle.tech/en/posts/blog/… @Google #Android

SafeCircle tokenomics is built to reward long-term holders: Tier 1 unlocks at 500k $SC with a 10% discount on all plans + early product news Tier 2 at 1M $SC adds 15% off, early access to updates, and selected beta features Tier 3 at 2M $SC gives 20% off, priority access, full exclusive features, and a private token-holder community. Learn more: safecircle.tech/token

Downloading your shield. Our AI models are hosted on @huggingface , and we’re grateful for their support. SafeCircle is getting closer.

🔒 50,000,000 $SC locked on RobinFlow — non-custodial and verifiable onchain on Robinhood Chain. Unlocks July 27, 2027. robinflow.app/t/0x8beF3368Af…

SafeCircle is now officially registered in Spain’s Intellectual Property Registry as a computer program. This matters because it protects our software, strengthens trust with partners and investors, and shows we’re building with a long-term, defensible foundation. We’re also in the process of registering across Europe. Big step forward for SafeCircle.

Dex paid

Updated CA, on X, github and website safecircle.tech/token

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.

SafeCircle $SC is live in Robinhood. 🚀 Protecting children online shouldn't mean spying on them. 🛡 SafeCircle uses AI to detect risky conversations, grooming attempts, and harmful content — without reading every message or invading your child's privacy. Built for families. Built for schools. Built for Europe's privacy standards. 🔗 safecircle.tech/token CA: 0x8bef3368af47c29746d13a6f1f3844dd685b90d2

@DogbillionMars Yes, soon more news

@SafeCircleAI are you guys relaunching?

Horizon Edge 4B is here. Our updated model is lean, fast, and built to push stronger on-device intelligence without the extra noise. Smarter performance, tighter efficiency, and a big step forward for our detection pipeline huggingface.co/safecircleai/h…

I’ve tried repeatedly to get in contact with the @Pumpfun team, and while they did respond, they declined to help. We launched the SC token in part to help fund the project, but with no fees and seeing other founders able to use creator fees to support their work, we have no choice but to discontinue the official SC token. We’ve been in talks with other founders about a relaunch. SC remains a foundational part of the project. More news soon

This is intrusive — treating young people like surveillance targets harms trust and privacy, and it’s not the way to keep them safe. We build privacy-first tools that protect children online without invasive monitoring, using consent, transparency, and minimal data processing to flag real risks while keeping family trust intact.

The Digital Services Act is helping make the online world safer for children. A new report highlights how the #DSA helps protect minors by addressing risks such as harmful content, cyberbullying, grooming and addictive platform design. Check the report: link.europa.eu/8xFMxK

SafeCircle AI Playground v1.0.1 is live! Fixed a bug where the on-device Horizon safety detection model would crash on longer conversations. Everything now runs smoothly, all analysis stays 100% on-device, zero data sent to servers. Update available now on GitHub github.com/safecircleia/a…

Horizon is running fully on-device in the SafeCircle AI Playground demo. That means offline chat, lower latency, and privacy-first inference without sending conversations to the cloud.

🔜

FYI: we’re still committed to open source. You can find all the models powering our detection pipeline on Hugging Face. huggingface.co/safecircleai

We improved the landing page to include new animations, check it out :)

AI Playground is almost on Google Play. Stay tuned.