Prakasa Windows CLI

Prakasa Windows CLI is a Windows environment configuration tool for the Prakasa decentralized P2P GPU inference network. It provides automatic detection and installation of WSL2, CUDA toolkit, and runtime dependencies required for participating in the Prakasa network. This project adopts modern C++ architecture and implements one-click deployment through NSIS-built Windows installer.

Prakasa — Decentralized P2P GPU Inference Network

Prakasa (Sanskrit: प्रकाश) — The decentralized "Light" of intelligence.

Prakasa is a decentralized, privacy-preserving P2P GPU inference middleware built on top of the open-source Parallax engine. It leverages the Nostr protocol for resilient, censorship-resistant orchestration and integrates the RIM economic system for trustless, real-time settlements and reputation.

About Prakasa

Prakasa is a high-performance, privacy-centric P2P GPU inference middleware that transforms idle GPU resources into a unified, resilient intelligence layer. Key features include:

Nostr-Powered Orchestration: Decentralized communication bus with no single point of failure
RIM Economic System: Atomic settlements with real-time micropayments per inference
Privacy & Security: End-to-end encryption, NIP-44/59 Gift Wrap for metadata protection
Built on Parallax: Evolution of the open-source Parallax engine for global P2P scale

Features

1. Environment Management

Automatic Check: Comprehensive checking of Windows environment, WSL2, NVIDIA drivers, and Prakasa runtime
One-Click Installation: Automatic installation and configuration of WSL2, CUDA toolkit, Prakasa dependencies
Smart Detection: Support for NVIDIA GPU detection and CUDA toolkit verification
System Requirements: Minimum support for RTX 3060 Ti, recommended RTX 4090 or higher configuration

2. Prakasa Network Integration

P2P Node Setup: Configure your Windows machine as a Prakasa compute provider
Nostr Integration: Automatic setup of Nostr relay connections for decentralized orchestration
RIM Wallet: Integrated settlement system for receiving compute payments
Real-time Feedback: Monitor installation progress and network connectivity status

3. WSL2 Integration

Command Execution: Complete WSL2 command execution support for Prakasa runtime
Real-time Output: Real-time stdout/stderr output for inference job monitoring
Interrupt Support: Support for Ctrl+C interruption of running compute tasks
Proxy Synchronization: Automatic synchronization of system proxy configuration for Nostr relay access

4. Configuration Management

Flexible Configuration: Support for multiple configurations including proxy, WSL distribution, installation sources, etc.
Dynamic Updates: Configuration changes automatically synchronized to related components
Reset Function: One-click reset of all configurations to default state
Secure Storage: Secure storage and access control for configuration files

5. Modern Architecture Design

Template Method Pattern: Unified command execution flow and error handling
CRTP Technology: Compile-time polymorphism with zero runtime overhead
Command Pattern: Extensible command registration and execution mechanism
Strategy Pattern: Flexible environment requirements and checking strategies
C++ Shell + Python Core: Windows native shell forwarding to Prakasa Python runtime

Architecture Design

Prakasa Windows CLI is a C++ native shell that manages the Prakasa Python runtime in WSL2. The architecture includes:

Command Line Interface (src/parallax/cli/): Unified command parsing with template method pattern
Environment Installer (src/parallax/environment/): Modular system for WSL2, CUDA, and Prakasa setup
Configuration Manager (src/parallax/config/): Dynamic configuration loading and validation
Utility Library (src/parallax/utils/): WSL command building, process management, GPU detection
Logging System (src/parallax/tinylog/): Asynchronous logging with rotation support

For detailed architecture analysis, see ARCHITECTURE_ANALYSIS.md.

System Requirements

Operating System: Windows 10 Version 2004+ or Windows 11 (x86_64, Administrator privileges required)

Hardware:

GPU: NVIDIA RTX 3060 Ti or higher (8GB+ VRAM, 24GB recommended)
RAM: 16GB minimum (32GB recommended)
Storage: 50GB+ available space
Network: Stable internet connection

Software (automatically installed except CUDA):

WSL2 + Ubuntu 24.04
NVIDIA Driver with CUDA 12.x support
CUDA Toolkit 12.8 or 12.9 (pre-install required)
Prakasa Runtime (Python-based)

Quick Start

1. Download and Install

Download the latest installer from the Release page:

Prakasa_Windows_Setup.exe

2. Environment Check

prakasa check

3. Environment Installation (if needed)

prakasa install

4. Configure Proxy (optional)

prakasa config proxy_url "http://127.0.0.1:7890"

# pip index mirror (optional, for faster package installation in China)
prakasa config pip_index_url "https://pypi.tuna.tsinghua.edu.cn/simple/"

5. Verify Installation

# Verify all environment components
prakasa check

# Start Prakasa compute node
prakasa run

# Join Prakasa network as provider
prakasa join

# Access chat interface (optional)
prakasa chat

Command Reference

`prakasa check`

Check environment requirements and Prakasa node readiness

prakasa check [--help|-h]

`prakasa install`

Install required environment components and Prakasa runtime

prakasa install [--help|-h]

`prakasa config`

Configuration management command

# View all configurations
prakasa config list

# Set configuration item
prakasa config <key> <value>

# Delete configuration item
prakasa config <key> ""

# Reset all configurations
prakasa config reset

# View help
prakasa config --help

`prakasa run`

Run Prakasa node (scheduler mode) directly in WSL

prakasa run [args...]

`prakasa join`

Join Prakasa P2P network as a compute provider

prakasa join [args...]

`prakasa chat`

Access Prakasa chat interface for testing inference

prakasa chat [args...]

`prakasa cmd`

Execute commands in WSL or Prakasa Python virtual environment

prakasa cmd [--venv] <command> [args...]

Command Descriptions:

run: Start Prakasa node in scheduler mode. Acts as a coordinator in the P2P network. You can pass any arguments supported by prakasa run command. Examples: prakasa run -m Qwen/Qwen3-0.6B, prakasa run --port 8080
join: Join Prakasa P2P network as a compute provider. Your GPU will be available for decentralized inference tasks. Examples: prakasa join -m Qwen/Qwen3-0.6B, prakasa join -s scheduler-addr
chat: Access chat interface for testing Prakasa inference capabilities. Examples: prakasa chat (local network), prakasa chat -s scheduler-addr (remote scheduler), prakasa chat --host 0.0.0.0 (allow external access). After launching, visit http://localhost:3002 in your browser.
cmd: Pass-through commands to WSL environment, supports --venv option to run in Prakasa project's Python virtual environment

Main Configuration Items:

proxy_url: Network proxy address (supports http, socks5, socks5h) - for Nostr relay access
pip_index_url: Python package index mirror URL (optional, for faster package installation in China)
wsl_linux_distro: WSL Linux distribution (default Ubuntu-24.04)
wsl_installer_url: WSL installer download URL
wsl_kernel_url: WSL2 kernel update package download URL
prakasa_git_repo_url: Prakasa project Git repository URL (default: https://github.com/hetu-project/prakasa.git)

Build Instructions

Development Environment

Visual Studio 2022: With C++ desktop development workload
CMake 3.20+: For building C++ projects
NSIS 3.08+: For creating Windows installer

Quick Build

Execute in project root directory:

cd src
mkdir build && cd build
cmake ../parallax -A x64
cmake --build . --config Release

Generated executable is located at: src/build/x64/Release/prakasa.exe

Create Installer

# 1. Create installation file directory
mkdir installer\FilesToInstall

# 2. Copy generated executable
copy src\build\x64\Release\prakasa.exe installer\FilesToInstall\

# 3. Build installer
call installer\build-nim-nozip.bat

Configuration

Prakasa is configured through configuration files, main configuration items include:

# WSL related configuration
wsl_installer_url=https://github.com/microsoft/WSL/releases/download/2.4.13/wsl.2.4.13.0.x64.msi
wsl_kernel_url=https://wslstorestorage.blob.core.windows.net/wslblob/wsl_update_x64.msi
wsl_linux_distro=Ubuntu-24.04

# Prakasa project configuration
prakasa_git_repo_url=https://github.com/hetu-project/prakasa.git

# Network proxy configuration (optional, for Nostr relay access)
proxy_url=http://127.0.0.1:7890

# Python package index mirror (optional, for faster package installation in China)
pip_index_url=https://pypi.tuna.tsinghua.edu.cn/simple/

Troubleshooting

Common Issues:

Environment Check Failed: Run with administrator privileges, verify Windows version and NVIDIA driver
WSL2 Installation Failed: Enable Virtual Machine Platform in Windows Features, check BIOS virtualization
CUDA Not Detected: Pre-install CUDA Toolkit 12.8 or 12.9, verify with nvidia-smi
Network Issues: Configure proxy if needed for Nostr relay access (prakasa config proxy_url "http://...")

Diagnostic Commands:

prakasa check              # Full environment check
prakasa config list        # View configuration
wsl --list --verbose       # WSL status
nvidia-smi                 # GPU status

Logs: prakasa.log in installation directory (C:\Program Files (x86)\Prakasa\)

Key Technical Features

Real-time Command Output: Live stdout/stderr with Ctrl+C interrupt support
Automatic Proxy Sync: Proxy configuration forwarded to WSL and Nostr relays
Smart GPU Detection: Automatic NVIDIA GPU and CUDA compatibility verification
Comprehensive Error Handling: Detailed diagnostics with recovery suggestions

Development

For development guidelines, build instructions, code style, and contribution workflow, see:

Developer Guide: AGENTS.md
Architecture Details: ARCHITECTURE_ANALYSIS.md

Directory Structure

prakasa_windows_cli/
├── src/prakasa/           # C++ source code
│   ├── cli/              # Command line interface
│   ├── environment/      # Environment installer
│   ├── config/           # Configuration manager
│   ├── utils/            # Utilities and WSL integration
│   └── tinylog/          # Logging system
├── docker/               # Cross-compilation support
├── installer/            # NSIS installer scripts
├── AGENTS.md            # Developer guide
├── ARCHITECTURE_ANALYSIS.md  # Technical architecture
└── README.md            # This file

Contributing

Contributions welcome! Please fork the repository, create a feature branch, and submit a Pull Request. For issues or questions, open an issue on GitHub.

License

See LICENSE file for details.

Note: Prakasa is a decentralized P2P GPU inference network built on the Parallax engine. It leverages Nostr protocol for resilient orchestration and RIM economic system for trustless settlements. Please ensure your hardware configuration meets the minimum requirements and that relevant drivers are correctly installed before participating in the network.

Additional Resources

Prakasa Core Repository: github.com/hetu-project/prakasa
Architecture Analysis: See ARCHITECTURE_ANALYSIS.md for detailed technical design
Developer Guide: See AGENTS.md for development guidelines
Nostr Protocol: github.com/nostr-protocol/nostr
Original Parallax: github.com/gradienthq/parallax

Name		Name	Last commit message	Last commit date
Latest commit History 28 Commits
.github/workflows		.github/workflows
installer		installer
src/parallax		src/parallax
.gitattributes		.gitattributes
.gitignore		.gitignore
AGENTS.md		AGENTS.md
ARCHITECTURE.md		ARCHITECTURE.md
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
README.md		README.md
USER_GUIDE.md		USER_GUIDE.md

Folders and files

Latest commit

History

Repository files navigation

Prakasa Windows CLI

Prakasa — Decentralized P2P GPU Inference Network

About Prakasa

Features

1. Environment Management

2. Prakasa Network Integration

3. WSL2 Integration

4. Configuration Management

5. Modern Architecture Design

Architecture Design

System Requirements

Quick Start

1. Download and Install

2. Environment Check

3. Environment Installation (if needed)

4. Configure Proxy (optional)

5. Verify Installation

Command Reference

prakasa check

prakasa install

prakasa config

prakasa run

prakasa join

prakasa chat

prakasa cmd

Build Instructions

Development Environment

Quick Build

Create Installer

Configuration

Troubleshooting

Key Technical Features

Development

Directory Structure

Contributing

License

Additional Resources

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 3

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

`prakasa check`

`prakasa install`

`prakasa config`

`prakasa run`

`prakasa join`

`prakasa chat`

`prakasa cmd`

Packages