Crynux

Crynux is a decentralized orchestration layer for edge AI, connecting a global network of devices to provide permissionless AI inference, fine-tuning, and training. Its production-ready AI cloud makes these capabilities available through APIs, so developers can build applications without managing local hardware. The network supports text, image, music, and video models, ranging from large models to quantized versions designed for edge computing.

Crynux uses its vssML consensus protocol, combining Zero-Knowledge Proofs and Verifiable Random Functions to verify computation and help resist Sybil attacks. Its blockchain-based system is designed to make transactions transparent, while distributing revenue to model token holders. Users can contribute data to training and earn rewards, with Federated Learning and Multi-Party Computation used to protect data.

The ecosystem offers several ways to participate: GPU owners can run nodes and earn CNX rewards, while others can stake CNX with node operators. Developers can use the API to build decentralized AI applications. Crynux also describes plans for AI workflows, model tokenization, multimodal support, and cross-chain assets.