The team behind Tether’s widely-used stablecoin, USDT, has introduced the QVAC SDK—an open-source software development kit designed to execute artificial intelligence applications directly on local devices without cloud connectivity.
Tether announced that this cross-platform toolkit functions seamlessly across iOS, Android, Windows, macOS, and Linux. Built upon QVAC Fabric—a derivative of llama.cpp—it ensures compatibility with the llama.cpp ecosystem for tasks like text generation, embeddings, and multimodal processing.
Developers can utilize a unified interface to access functionalities such as text completion, embeddings, vision recognition, optical character recognition, text-to-speech, speech-to-text, and translation. The toolkit’s peer-to-peer architecture, powered by the Holepunch stack, facilitates decentralized model distribution and delegated inference, with plans for future updates including swarms for decentralized training, fine-tuning, and inference.
Paolo Ardoino, CEO of Tether, commented on the shift in technological needs: “As we approach a time when billions of humans coexist with trillions of AI agents, the current centralized model cannot scale effectively.” He emphasized that physical limitations like speed-of-light latency, single points of failure, and control concentration hinder centralized AI’s future. “QVAC is crafted for this emerging world,” he added. “It lays the groundwork for what we call the Stable Intelligence Era.”
Originating from Tether Data’s AI research initiative focused on developing decentralized and adaptive intelligence systems, QVAC represents a significant commitment by the company to expand its open-source ecosystem in forthcoming months, including specialized toolkits for robotics and brain-computer interfaces.
This launch marks an important strategic expansion for Tether beyond its primary stablecoin operations. The company aims to challenge centralized AI providers by offering privacy-centric alternatives that process data locally, eliminating the need to transmit it to cloud servers.