
KeryxKRX
A GPU-mined BlockDAG that integrates decentralized AI inference through Proof-of-Model and Optimistic Proof of Inference.

Permissionless AI inference, fine-tuning and training powered by a global network of edge devices.
PROJECT OVERVIEW
Crynux is a decentralized orchestration layer for artificial intelligence workloads. It brings together computing resources from home computers, servers and other edge devices, allowing applications to access distributed GPU capacity for AI inference, image generation, fine-tuning and training.
Developers can connect applications through the Crynux Bridge, OpenAI-compatible APIs and software development kits. Compatible node operators can install the open-source Crynux Node software, contribute computing resources and receive CNX rewards for successfully completed workloads.
The network uses blockchain-based task coordination, staking, node-performance scoring and cryptographic verification mechanisms intended to discourage incorrect or malicious computation. Its vssML design combines Verifiable Random Functions and Zero-Knowledge Proofs to selectively validate AI tasks while reducing the additional computation required for consensus.
CNX is used for network fees, AI task payments, node staking, delegated staking and rewards. Crynux describes a multi-chain architecture in which its Layer 2 networks can operate across EVM-compatible blockchain ecosystems.
Network Type: Decentralized edge-AI compute network
Architecture: Multi-chain EVM-compatible Layer 2 architecture
Supported AI Workloads:
• Large language model inference
• Vision-language model tasks
• Stable Diffusion and image generation
• Model fine-tuning and training
Developer Access:
• OpenAI-compatible Crynux Bridge API
• Python SDK
• JavaScript SDK
• Direct blockchain and Relay integration
Node Software:
• Open source
• Windows support
• Apple Silicon Mac support
• Docker support
• Cloud GPU support
Windows Hardware Requirement:
• NVIDIA GPU
• Minimum 8 GB VRAM
Mac Hardware Requirement:
• Apple Silicon M1 or newer
Task Coordination:
• Blockchain smart contracts
• Crynux Relay for off-chain task data and results
• Node staking
• Quality of Service scoring
• Task pricing and queue prioritisation
Verification Design:
• vssML
• Verifiable Random Functions
• Zero-Knowledge Proofs
• Selective task validation
• Similarity-based output verification
• Slashing for identified malicious behaviour
Supported Crynux L2 Environments:
• Dymension
• NEAR
Crynux’s official website describes Lithium as its first mainnet and currently invites users to join the mainnet. However, some sections of the technical documentation still reference testnet RPC endpoints, test CNX and non-transferable Node Credits.
Users should verify the current network, bridge and token configuration through the official Crynux Portal and documentation before transferring CNX, staking tokens or operating a node.
PROJECT PULSE
A new project profile is now available in the Chainquiry directory.
Chainquiry project profile ↗Get important project changes by email. No account is required for a single-project watch.
Want a full watchlist and notification history? Create a free account.