Tether unveils QVAC SDK for local AI on all devices

Tether launched QVAC SDK, an open-source toolkit for running AI apps locally on devices, supporting text, speech, vision, and translation, with peer-to-peer model sharing and plans for decentralized training.

Tether has launched QVAC SDK, an open-source software development kit that enables artificial intelligence applications to run locally on a wide range of devices, including smartphones, laptops, desktops, and servers. The SDK is compatible with major operating systems such as iOS, Android, Windows, macOS, and Linux. Built on QVAC Fabric, a fork of llama.cpp, QVAC SDK supports text generation, embeddings, multimodal workloads, speech processing, visual recognition, and translation. It integrates specialized engines like whisper.cpp, Parakeet, and Bergamot, all accessible through a unified interface. By allowing AI models to be built, trained, and deployed directly on devices, QVAC SDK eliminates the need for cloud servers. It utilizes the Holepunch protocol for peer-to-peer model distribution and delegated inference. Tether plans to expand the toolkit with decentralized training, fine-tuning, and modules for robotics and brain-computer interfaces, aiming to enhance privacy, reduce latency, and ensure AI services remain functional even with limited connectivity.