Decentralized Training
Train large language models across thousands of consumer GPUs using our DiLoCo protocol. 500x less communication than standard distributed training.
Train models together. Earn NEURO. Own collective intelligence.

Get up and running in 3 minutes:
# Install NeuroShard
pip install neuroshard-ai
# Run a node (GPU recommended)
neuroshard --token YOUR_WALLET_TOKENYour node starts on port 8000 (HTTP dashboard) and 9000 (gRPC). Get your token at neuroshard.com/signup.