Google AI Edge

Google AI Edge is Google’s technology stack for building and deploying AI and machine learning directly on devices instead of relying entirely on cloud processing. Its ecosystem includes MediaPipe for prebuilt AI tasks such as face mesh, object detection, and background effects; LiteRT-LM for running large language models across Android, iOS, web, and embedded devices; and LiteRT for deploying hardware-accelerated custom models. Developers can work with models from frameworks including PyTorch, JAX, TensorFlow, and Keras and target hardware such as CPUs, GPUs, and NPUs. Google also provides tools such as Model Explorer, AI Edge Quantizer, and AI Edge Portal. The platform is designed for developers building privacy-conscious, low-latency, offline-capable, or hardware-optimized AI experiences.