Even Ollama says this local AI inference is cool – Nexa SDK for NPU

Show HN: Even Ollama says this local AI inference is cool – Nexa SDK for NPU

Even Ollama says this local AI inference is cool – Nexa SDK for NPU operates in show hn launch.

Profile

We built Nexa SDK to make local AI development simple.• Run any model—text, vision, audio, speech, image-gen—fully on-device• Works on NPU, GPU, or CPU with built-in support for Qualcomm & Apple NPUs, GGUF, and Apple MLX• Supports SOTA models like Gemma-3n, PaddleOCR, Qwen3, Parakeet v3, Phi-4, and our own OmniNeural-4BNo fragmented runtimes, no complex setup—just one lightweight toolkit from prototype to production.We were recently featured at Snapdragon Summit 2025 and are working with Qualcomm, AMD, and Intel to make on-device AI more accessible.Would love feedback from the HN community—performance numbers, build experiences, feature requests, anything.We’re also celebrating our Product H

Even Ollama says this local AI inference is cool – Nexa SDK for NPU is positioned in show hn launch, a segment that rewards a clear, focused offering.

Its current recorded stage is Show HN launch.

Website: sdk.nexa.ai