A peer-to-peer compute protocol for decentralized AI development at scale. This protocol implementation powered distributed training runs including synthetic-2 and intellect-2.
This repository contains the core infrastructure for coordinating compute resources across a distributed network, including workers, validators, and the orchestration layer.
- System Architecture
- Getting Started
- Documentation
- Frequently Asked Questions
- Community
- Contributing
- Security
- License
The protocol follows a modular architecture designed for decentralized AI compute:
- Smart Contracts: Ethereum-based contracts manage the protocol's economic layer
- Discovery Service: Enables secure peer discovery and metadata sharing
- Orchestrator: Coordinates compute jobs across worker nodes
- Validator Network: Ensures quality through random challenges
- Worker Nodes: Execute AI workloads in secure containers
- Linux operating system
- CUDA-capable GPU(s) for worker operations
- Docker (version 28.1.1 or later) and Docker Compose (version v2.35.1 or later)
For complete setup instructions, refer to our Development Setup Guide.
You can install the latest worker CLI using:
curl -sSL https://raw.githubusercontent.com/PrimeIntellect-ai/protocol/main/crates/worker/scripts/install.sh | bash
This can also be used to upgrade the current installation to the latest release.
For the latest dev build use:
curl -sSL https://raw.githubusercontent.com/PrimeIntellect-ai/protocol/develop/crates/worker/scripts/install.sh | bash -s -- --dev
- Development Setup Guide - Detailed installation and environment setup instructions
- Usage Guide - Instructions for dispatching tasks, monitoring, and system management
A: This is a peer-to-peer compute protocol implementation for decentralized AI development at scale. It provides infrastructure for coordinating compute resources across a distributed network through workers, validators, and an orchestration layer. This protocol powered distributed training runs including synthetic-2 and intellect-2.
A: Yes, you can run the protocol locally for development and testing purposes.
A: We have provided an .env.example file with the required variables.
A: We're actively developing our security practices for private key management.
A: We will be providing detailed documentation on how to secure workers with firewalls for ingress / egress
A: For the current development phase, minimal testnet ETH is sufficient. Detailed information on setting up the worker will follow.
We welcome contributions! Please see our Contributing Guidelines.
See SECURITY.md for security policies and reporting vulnerabilities.
