Show HN: Free On-Device AI SDK to Run PyTorch on Mobile NPUs (Open Source) (github.com)

Hello HN,

I’m Yeonseok, CEO of ZETIC.ai (ex-Qualcomm AI Research).

Deploying AI models to mobile NPUs usually requires deep hardware knowledge and months of tuning. Today, we are changing that.

We are opening our On-device AI SDK for free right now.

We’ve released a GitHub repository that gives you direct access to full NPU acceleration on heterogeneous devices (Android/iOS). You don't need GPU Cloud infrastructure anymore.

What you can do with this SDK (GitHub): Full NPU Speed: We unlock raw NPU power for heterogeneous devices globally. No more CPU lag. Bring Your Own Model: Don't just run our demos. You can upload your custom model and get an optimized, hardware-accelerated SDK in hours. Production-Ready Source Code: The repo includes full source code for On-device AI Agents, 40+ Language Translations, and SOTA Object Detection examples to get you started immediately. We want to democratize access to NPU computing. The GitHub repo is ready-to-use. I’d appreciate it if you could check the code and give it a Star if you find it useful.

Repo: https://github.com/zetic-ai/ZETIC_MLange_apps

Let me know if you have any questions about the NPU implementation!