
google-ai-edge
5 Tools
Google AI Edge is a Google team building the end-to-end stack for on-device machine learning and generative AI, enabling developers to run models locally across mobile, web, and embedded systems without cloud dependency. The team's lineage traces to TensorFlow Lite, which launched in 2017 and reached 2.7 billion devices; in September 2024, the runtime was rebranded to reflect its evolution into a multi-framework edge AI platform supporting PyTorch, JAX, Keras, and TensorFlow alike.
Tools
Mobile app that runs open models like Gemma 4 fully on-device for offline chat, image analysis, and voice transcription.
Google's cross-platform ML framework for streaming media, with face, hand, and pose tracking plus audio pipelines.
Google framework for running large language models on phones, wearables, and IoT devices with GPU and NPU support.
Google's on-device ML runtime, successor to TensorFlow Lite, running models on mobile, embedded, and edge hardware.
Command-line tool for the LiteRT on-device runtime to convert, quantize, run, and benchmark models on CPU, GPU, and NPU.