The cross-platform toolkit works unchanged across iOS, Android, Windows, macOS, and Linux environments, the company said. The SDK is built on QVAC Fabric, a fork of llama.cpp, providing compatibility with the llama.cpp model ecosystem for text generation, embeddings, and multimodal workloads.
Through a single interface, developers can access text completion, embeddings, vision, optical character recognition, text-to-speech, speech-to-text, and translation capabilities. The toolkit's peer-to-peer functionality, powered by the Holepunch stack, enables decentralized model distribution and delegated inference. The company plans to add peer-to-peer swarms for decentralized training, fine-tuning, and inference in future updates.
"The world is approaching a moment where billions of humans share the planet with billions of autonomous machines and trillions of AI agents," Tether CEO Paolo Ardoino said, in a statement. "The current model, routing every decision through a centralized server, won't scale to meet that reality."
"The laws of physics alone make centralized AI a dead end: speed-of-light latency, single points of failure, and concentration of control are features of a system designed for a smaller world," he added. "QVAC is built for the world that’s coming. It is the building block of the Stable Intelligence Era.”
The launch marks a significant strategic expansion for Tether beyond its core stablecoin business, positioning the company to compete with centralized AI providers by offering privacy-focused alternatives that process data locally rather than transmitting it to cloud servers.
