Aura SDK v0.2.1 Released: Library Refactoring & Automated Workflows
We are proud to release Aura SDK v0.2.1. This major update transforms the project from a procedural CLI utility into a clean, reusable library. Programmatic integrations now feature closure-based callbacks, and our new CI/CD workflow automates Windows-ARM64 compiles, testing, and release publishing.
Aura SDK Repository Successfully Transferred to Official Organization
We are pleased to announce that our flagship Edge AI inference engine, Aura SDK, has been officially transferred to our official GitHub organization: Artificial-Intelligence-Lab-ARL. This transition enables improved governance, developer access, and a unified workspace for upcoming compiler tools and model quantization libraries.
Aura SDK v0.1.0 Released with Native Hexagon NPU Support
Our first public developer release of Aura SDK is live. Version Alpha delivers low-latency model graph compilation and direct zero-copy weight mapping for Snapdragon X Series processors, bypassing cloud roundtrip latencies to execute LLMs at up to 35 tokens per second locally.
Joint Engineering Labs Program Commences
To accelerate on-device AI integration in enterprise environments, we have launched the Joint Engineering Labs. Selected developer partners gain priority support from our system architects to compile custom model architectures and optimize execution enclaves.