Fixed #

• Native binaries now bundled directly in the pub.dev package (no external download needed)
• Auto-detection of bundled binaries for zero-config consumption
• Updated dependency to runanywhere ^0.17.0 (adds InferenceFramework.genie)

Changed #

• Updated license to reflect Qualcomm partner distribution terms
• Native binaries automatically downloaded at build time from GitHub releases
• Fixed copyright year

Initial Release #

• Qualcomm Genie NPU backend integration for RunAnywhere Flutter SDK
• Support for Snapdragon 8 Gen 2+ devices
• ARM64 (arm64-v8a) architecture support
• Automatic binary download mechanism for native libraries
• Integration with RunAnywhere Flutter SDK v0.16.0+

Features #

• LLM inference using Qualcomm Snapdragon Neural Processing Unit
• Optimized performance for on-device AI models
• Seamless plugin registration with Flutter

Technical Details #

• Binary libraries automatically downloaded from GitHub releases
• Includes Qualcomm QNN runtime libraries
• Native library size: ~65MB compressed, ~178MB uncompressed