GitHub - AIPowerGrid/grid-media-worker

╔═══════════════════════════════════════════════════════════════╗
║                                                               ║
║            █████╗ ██╗██████╗  ██████╗                         ║
║           ██╔══██╗██║██╔══██╗██╔════╝                         ║
║           ███████║██║██████╔╝██║  ███╗                        ║
║           ██╔══██║██║██╔═══╝ ██║   ██║                        ║
║           ██║  ██║██║██║     ╚██████╔╝                        ║
║           ╚═╝  ╚═╝╚═╝╚═╝      ╚═════╝                         ║
║                                                               ║
║        Turn Your GPU Into a Money-Making Machine              ║
║        and Help Power the Community                           ║
║                                                               ║
╚═══════════════════════════════════════════════════════════════╝

🎨 Generate AI art and videos while earning AIPG tokens!

⏱️ Heads up: The very first build takes 45–60 minutes (Docker images, model cache, desktop app). Let it run once; every restart after that is usually under two minutes.

=======

Docker Desktop (Free version is fine) (https://www.docker.com/products/docker-desktop/)
Node.js (https://nodejs.org/en/download)
An API key from AI Power Grid
An API key from Huggingface (once logged in)
An API key from Civitai (once logged in)

🚀 Quick Start

Get up and running in under 2 minutes - everything happens automatically!

🔑 Get your API key from dashboard.aipowergrid.io
⚙️ Copy env.example to .env and add your API key
▶️ Right-click and "Run as Administrator" the start script:
- Windows: start-worker.bat (right-click → Run as Administrator)
- Mac/Linux: sudo ./start-worker.sh
⏳ Plan 45–60 minutes for the very first run. Docker pulls, dependency installs, model downloads, and the desktop app build all happen automatically—later launches are fast.
✨ That's it! The script automatically:
- ✅ Checks all requirements (disk space, Docker, etc.)
- ✅ Starts Docker containers
- ✅ Downloads models listed in .env during image build
- ✅ Builds the desktop app automatically
- ✅ Cleans up stale Docker volumes so upgrades stay reliable
🖥️ Desktop app is built automatically and available in management-ui-nextjs/dist/
🎨 Open the desktop app OR http://localhost:5000 in your browser
💰 Click "Start Hosting" → You're earning!

💡 Desktop App: The Electron app is built automatically during Docker build - no Node.js needed on your host! Find it in management-ui-nextjs/dist/ after building.

⚡ Pro Tip: Everything happens automatically - Docker builds the desktop app, downloads models, and sets everything up. Just run the script and you're done!

💰 Why Run This Worker?

Turn Idle Time Into Income 💵

Your GPU sits idle most of the time. Why not put it to work? You earn AIPG tokens just by having the worker running - even while you sleep!

Simple Setup, Zero Hassle ⚡

One-click start - no complex configuration needed
Web UI - manage everything from your browser
Automatic updates - always running the latest version
24/7 operation - set it and forget it

Power the Future of AI 🌐

Join thousands of GPU owners helping build a decentralized AI network. Make AI generation accessible to everyone while earning rewards.

🎁 What You Get

Benefit	What It Means
💵 Earn Tokens	Get paid in AIPG just for running the worker
🎨 Free AI Access	Use powerful AI models without paying
🔒 Secure	Your API keys stay private, never shared
📊 Real-time Dashboard	Track earnings and performance live
🌐 Decentralized	Support the open AI network
⚡ Easy Management	Web UI and desktop app make everything simple
🖥️ Desktop App	Native desktop application built automatically during Docker build

🖥️ Desktop App (Electron)

The Electron desktop app is automatically built during Docker image build! No manual setup required - it's ready to use right after docker-compose build.

✨ Automatic Build (Recommended)

When you run docker-compose build, the Electron desktop app is built automatically:

# Build everything including Electron app
docker-compose build

# The built app is available at:
# management-ui-nextjs/dist/

What you get:

✅ Built automatically - No manual steps needed
✅ Ready to use - Available in management-ui-nextjs/dist/ after build
✅ Platform-specific - Built for your container platform (Linux)
✅ Skip if needed - Set BUILD_ELECTRON=false to skip

To skip Electron build:

# Option 1: Environment variable
BUILD_ELECTRON=false docker-compose build

# Option 2: Add to .env file
echo "BUILD_ELECTRON=false" >> .env

📦 Finding Your Built App

After building, check the dist/ directory:

Linux: dist/linux-unpacked/ or dist/*.AppImage or dist/*.deb
Windows/Mac: Build on those platforms or use manual build scripts below

🛠️ Manual Build (Optional)

If you need to build manually outside Docker or for a different platform:

Prerequisites:

Node.js 18+ installed
Docker containers running (app connects to http://localhost:5000)

Quick build:

cd management-ui-nextjs
npm install
npm run electron:build    # Production build
# OR
npm run electron:pack     # Portable build (Windows)

Development mode:

npm run electron:dev      # Auto-reloads on code changes

🚀 Launching the App

From Docker build:

Navigate to management-ui-nextjs/dist/
Run the executable for your platform
App connects to http://localhost:5000 automatically

Features:

✅ Native window controls and system integration
✅ No browser needed - standalone application
✅ Better performance and offline capability
✅ Same functionality as web interface

📋 Requirements

Software (install once)

✅ Docker Desktop – Download for Windows / macOS (Linux users install Docker Engine from your distro repos). Docker runs ComfyUI, Python, and every dependency in containers, so you never touch those directly.
✅ Node.js 18+ – Download installers. We use Node to build the AI Power Grid desktop app automatically.

💡 The start scripts verify both apps are installed/running and walk you through fixes if something is missing.

Hardware & Space (automatically checked)

💾 50GB+ free disk space (for Docker images + models)
🖥️ NVIDIA GPU with 6GB+ VRAM (or comparable AMD ROCm card)
🧠 8GB+ system RAM recommended

If a check fails, the script pauses with a friendly message so you can address it and re-run.

📖 Detailed Setup Guide

Step 1: Get Your API Key 🔑

Visit dashboard.aipowergrid.io
Sign up (it's free!)
Go to "API Keys" → Generate new key
Copy the key (starts with aipg_)

Step 2: Configure ⚙️

Windows:

copy env.example .env
notepad .env

Mac/Linux:

cp env.example .env
nano .env

Add these two lines (replace with your actual values):

GRID_API_KEY=aipg_your_key_here
GRID_WORKER_NAME=YourName.YourWalletAddress

📝 Worker Name Format: YourName.YourWalletAddress
Example: JohnDoe.0x1234567890abcdef1234567890abcdef12345678

Step 3: Start the Worker ▶️

Windows:

Double-click start-worker.bat

Mac/Linux:

Run: ./start-worker.sh
Or: chmod +x start-worker.sh && ./start-worker.sh

The script automatically:

✅ Checks disk space (50GB minimum)
✅ Verifies Docker is installed
✅ Ensures Docker is running
✅ Validates your configuration
✅ Builds the desktop app (if Node.js is installed)
✅ Creates a desktop shortcut for easy access

Step 4: Select Models 🎨

🔗 Blockchain Model Registry: All models are registered and validated through the ModelVault smart contract on Base Mainnet. This ensures authenticity, proper attribution, and trustless model discovery.

Option A: Use the Desktop App (Recommended - Created Automatically!)

Look for "AI Power Grid Manager" shortcut on your desktop
Double-click to launch (no browser needed!)
Browse available models (filter by your GPU's VRAM)
- All models are verified on the blockchain ✅
- Only registered models can be downloaded and hosted
Click Download for models you want to host
Wait for downloads to complete
Click "Start Hosting" → You're earning! 🎉

Option B: Use the Web Interface

Open http://localhost:5000 in your browser
Browse available models (filter by your GPU's VRAM)
- All models are verified on the blockchain ✅
- Only registered models can be downloaded and hosted
Click Download for models you want to host
Wait for downloads to complete
Click "Start Hosting" → You're earning! 🎉

💡 Note: The desktop app is automatically built and a shortcut is created when you run the start script. If the shortcut wasn't created, you can still use the web interface!

🔐 Security: All model information comes from the blockchain - no centralized server controls which models are available. This ensures transparency and prevents censorship.

📊 Monitor Your Earnings

Track your progress in real-time:

🖥️ Desktop App: Desktop shortcut created automatically! Launch "AI Power Grid Manager" from your desktop (no browser needed!)
🌐 Dashboard: dashboard.aipowergrid.io - View earnings, stats, and history
💻 Web UI: http://localhost:5000 - Manage models and monitor jobs in your browser
📝 View Logs: docker-compose logs -f - See detailed activity

🛠️ Troubleshooting

Worker Won't Start ❌

Check these first:

✅ Docker Desktop is running (look for whale icon in system tray)
✅ .env file exists with your API key
✅ Check logs: docker-compose logs for specific errors

Common fixes:

Restart Docker Desktop
Verify port 5000 or 8188 isn't already in use
Ensure you have enough disk space (50GB+)

Not Receiving Jobs? 🔍

Verify these:

✅ Models show "Hosting" status (green) in the UI
✅ API key is valid at dashboard.aipowergrid.io
✅ Worker name format is correct: Name.WalletAddress
✅ Worker is online (check dashboard)

Still not working?

Check logs: docker-compose logs -f
Verify your GPU meets model requirements
Ensure models finished downloading completely

GPU Not Detected? 🎮

For NVIDIA GPUs:

Install NVIDIA Container Toolkit
Run nvidia-smi to verify GPU is detected
Restart Docker Desktop
Check logs: docker-compose logs -f

For AMD GPUs:

Ensure ROCm is properly installed
Check Docker supports your GPU model

TypeScript/Build Errors? 🔧

If you see errors about missing modules or TypeScript types when working with the management UI:

cd management-ui-nextjs
npm install

This installs all required dependencies and type definitions.

Need More Help? 💬

📝 Logs: docker-compose logs -f - See what's happening
💬 Discord - Get help from the community
📚 Documentation - Full technical docs

🎮 GPU Requirements

Model Type	VRAM Needed	What It Does
SD 1.5	6GB	General images, fastest
SDXL	8GB	High-resolution images
Flux	12GB	Advanced, high-quality images
Video (5B)	16GB	Short videos (5-10 seconds)
Video (14B)	32GB	High-quality videos

💡 New to GPU hosting? Start with SD 1.5 models - they're the most popular and work great on entry-level GPUs!

🔄 Common Commands

Update to latest version:

git pull
docker-compose down
docker-compose build
docker-compose up -d

Stop the worker:

docker-compose down

Start the worker:

docker-compose up -d

View live logs:

docker-compose logs -f

Restart after changes:

docker-compose restart

⚙️ Advanced Configuration

Faster Model Downloads 🚀

Add these to your .env for faster downloads:

HUGGING_FACE_API_KEY=your_hf_token
CIVITAI_API_KEY=your_civitai_token

💡 These API keys are optional but can significantly speed up model downloads from Hugging Face and Civitai.

Download Models During Build 🏗️

Models listed in your .env file are automatically downloaded during the Docker image build!

This means:

✅ Models are ready immediately when containers start (no waiting at runtime)
✅ Faster container startup times
✅ Models are baked into the image for consistent deployments

How it works:

Add models to your .env file:

GRID_MODEL=FLUX.1-dev,SDXL,SD-1.5
# OR
WORKFLOW_FILE=flux.1_krea_dev.json,sdxl1.json

Rebuild the Docker image:

docker-compose build --no-cache comfy-bridge

Models will be downloaded during the build process

Note: If you add models to .env after the image is built, they'll be downloaded at runtime when the container starts. To download during build, rebuild the image with the updated .env file.

Build-time vs Runtime downloads:

Build-time: Models are in the Docker image, faster container startup
Runtime: Models are downloaded when container starts (if not in image or .env changed)
Both methods work - choose based on your workflow!

Blockchain Model Registry 🔗

All models are registered and validated through the ModelVault smart contract on Base Mainnet.

The blockchain is the single source of truth for:

✅ Model discovery and availability
✅ Model authenticity and verification
✅ Download URLs and file hashes
✅ Model constraints (steps, CFG, samplers)

Configuration (already set by default):

MODELVAULT_ENABLED=true                        # Enable blockchain registry
MODELVAULT_RPC_URL=https://mainnet.base.org    # Base Mainnet RPC
MODELVAULT_CONTRACT=0x79F39f2a0eA476f53994812e6a8f3C8CFe08c609  # Contract address

Benefits:

🔐 Trustless: No central authority controls which models are available
✅ Verified: All models are registered on-chain with cryptographic hashes
🌐 Transparent: Anyone can verify model registration and details
🚫 Censorship-resistant: No single entity can remove models from the registry

Wallet Connect Configuration 🔗

The Management UI supports Web3 wallet connections for blockchain features. A default WalletConnect project ID is provided, but you can get your own free project ID:

Visit https://cloud.walletconnect.com
Create a free account and new project

Add to your .env:

NEXT_PUBLIC_WALLETCONNECT_PROJECT_ID=your_project_id

Custom Ports 🔌

Edit docker-compose.yml to change ports if needed:

Default: Management UI on port 5000, ComfyUI on port 8188
Change if these ports conflict with other services

🏗️ How It Works

This worker runs ComfyUI (powerful AI generation engine) and connects it to the AI Power Grid network. You earn AIPG tokens just by keeping the worker running and hosting models!

Components:

🎨 ComfyUI - The AI generation engine that processes jobs
🌉 Bridge - Connects your worker to the AI Power Grid network
💻 Management UI - Web interface for easy model and job management

The Flow:

Your worker connects to the network
Jobs come in and your GPU processes them
You earn AIPG tokens for keeping the worker online
Payments are made every hour on the 20 minute mark! 💰

📚 Additional Resources

📖 Full Documentation - Complete technical docs
💬 Discord Community - Get help, share tips
📊 Dashboard - Track earnings and manage account
🐛 Report Issues - Found a bug?

❓ Frequently Asked Questions

💰 Earning & Payments

Q: How much can I earn?
A: Earnings depend on your GPU, models hosted, and network demand. Higher-end GPUs hosting popular models typically earn more. Check the dashboard for current rates.

Q: How do I get paid?
A: You earn AIPG tokens automatically just by having the worker running. No need to process jobs - just keep it online!

Q: When do I get paid?
A: Payments are made automatically every hour on the 20 minute mark (e.g., 1:20, 2:20, 3:20). Check your dashboard to see your earnings.

Q: Can I run this on multiple GPUs?
A: Yes! Each GPU can run its own worker instance. Just use different worker names and ensure each has enough resources.

🎮 GPU & Hardware

Q: Do I need a high-end GPU?
A: No! Entry-level GPUs (6GB VRAM) can run SD 1.5 models. Higher-end GPUs can run more advanced models and earn more.

Q: Can I use my GPU while the worker runs?
A: Yes, but it may slow down both. The worker uses GPU resources, so gaming or other GPU-intensive tasks may impact performance.

Q: Will this damage my GPU?
A: No. The worker runs at normal operating temperatures. Modern GPUs are designed for 24/7 operation. Monitor temperatures if concerned.

Q: Can I use an AMD GPU?
A: Yes! AMD GPUs with ROCm support work. Setup may differ slightly - check the documentation for AMD-specific instructions.

⚙️ Setup & Configuration

Q: Do I need technical knowledge?
A: No! The one-click scripts handle everything. Basic computer skills are enough.

Q: How long does setup take?
A: About 2-5 minutes for initial setup, plus model download time (varies by model size and internet speed).

Q: Can I change models later?
A: Yes! Use the web UI at http://localhost:5000 to download new models or stop hosting current ones anytime.

Q: Do I need to keep my computer on?
A: Yes, the worker needs your computer running. Many users run it 24/7 on dedicated machines or when idle.

🔒 Security & Privacy

Q: Is my API key safe?
A: Yes! Your API key stays on your machine and is never shared. Only you have access to it.

Q: Can others access my GPU?
A: No. Only jobs from the AI Power Grid network are processed. You control which models to host.

Q: What data is collected?
A: Only worker status and uptime metrics (for payments). No personal data or generated content is stored.

🐛 Troubleshooting

Q: Worker won't start
A: Check Docker is running, .env file exists, and you have enough disk space. Run docker-compose logs for details.

Q: Not receiving jobs
A: Verify models show "Hosting" status (green), API key is valid, and worker name format is correct.

Q: Jobs failing
A: Check GPU has enough VRAM for the model, logs for errors, and ensure models downloaded completely.

Q: Slow performance
A: Ensure no other GPU-intensive apps are running, check GPU temperatures aren't throttling, and verify Docker has GPU access.

🌐 Network & Connectivity

Q: Do I need fast internet?
A: Moderate speed is fine. Faster internet helps with model downloads and uploading results faster.

Q: Can I run this offline?
A: No, you need internet to receive jobs and upload results. The worker connects to the AI Power Grid network.

Q: What ports are used?
A: Port 5000 (Management UI) and 8188 (ComfyUI). These can be changed in docker-compose.yml if needed.

📊 Models & Performance

Q: Which models should I host?
A: Start with SD 1.5 (most popular). Then try SDXL or Flux if your GPU supports it. Video models require more VRAM.

Q: How many models can I host?
A: As many as your disk space allows. Each model needs to be downloaded and stored. Start with 2-3 popular models.

Q: Can I test models before hosting?
A: Yes! Use the ComfyUI interface at http://localhost:8188 to test models before enabling hosting.

Q: Why are some models not available?
A: Models may require more VRAM than your GPU has, or may not be supported yet. Check GPU requirements table above.

Made with ❤️ by ameli0x and half

💬 Discord | 📚 Documentation

Ready to turn your GPU into a money-making machine? 🚀

AI Power Grid for the API
ComfyUI for the local image generation backend

Name		Name	Last commit message	Last commit date
Latest commit History 302 Commits
comfy_bridge		comfy_bridge
management-ui-nextjs		management-ui-nextjs
management-ui		management-ui
public		public
scripts		scripts
tests		tests
workflows		workflows
.coverage		.coverage
.dockerignore		.dockerignore
.gitignore		.gitignore
ARCHITECTURE.md		ARCHITECTURE.md
BLOCKCHAIN.md		BLOCKCHAIN.md
FIX_RECIPE_SYNC.md		FIX_RECIPE_SYNC.md
LTXV_REGISTRATION.md		LTXV_REGISTRATION.md
README.md		README.md
RECIPESVAULT_NOT_REGISTERED.md		RECIPESVAULT_NOT_REGISTERED.md
RECIPESVAULT_TROUBLESHOOTING.md		RECIPESVAULT_TROUBLESHOOTING.md
RECIPE_REGISTRATION_FAILED.md		RECIPE_REGISTRATION_FAILED.md
RECIPE_REGISTRATION_SUMMARY.md		RECIPE_REGISTRATION_SUMMARY.md
REGISTER_RECIPE_GUIDE.md		REGISTER_RECIPE_GUIDE.md
build-electron-app.bat		build-electron-app.bat
build-electron-app.sh		build-electron-app.sh
docker-compose.yml		docker-compose.yml
docker-entrypoint.sh		docker-entrypoint.sh
dockerfile		dockerfile
download_models_from_chain.py		download_models_from_chain.py
downloads_api.py		downloads_api.py
env.example		env.example
get_gpu_info.py		get_gpu_info.py
gpu_info_api.py		gpu_info_api.py
install-optional-deps.sh		install-optional-deps.sh
model_catalog.py		model_catalog.py
model_manager.py		model_manager.py
package-lock.json		package-lock.json
package.json		package.json
pyproject.toml		pyproject.toml
requirements.txt		requirements.txt
start-worker.bat		start-worker.bat
start-worker.sh		start-worker.sh

Folders and files

Latest commit

History

Repository files navigation

🚀 Quick Start

💰 Why Run This Worker?

Turn Idle Time Into Income 💵

Simple Setup, Zero Hassle ⚡

Power the Future of AI 🌐

🎁 What You Get

🖥️ Desktop App (Electron)

✨ Automatic Build (Recommended)

📦 Finding Your Built App

🛠️ Manual Build (Optional)

🚀 Launching the App

📋 Requirements

Software (install once)

Hardware & Space (automatically checked)

📖 Detailed Setup Guide

Step 1: Get Your API Key 🔑

Step 2: Configure ⚙️

Step 3: Start the Worker ▶️

Step 4: Select Models 🎨

📊 Monitor Your Earnings

🛠️ Troubleshooting

Worker Won't Start ❌

Not Receiving Jobs? 🔍

GPU Not Detected? 🎮

TypeScript/Build Errors? 🔧

Need More Help? 💬

🎮 GPU Requirements

🔄 Common Commands

⚙️ Advanced Configuration

Faster Model Downloads 🚀

Download Models During Build 🏗️

Blockchain Model Registry 🔗

Wallet Connect Configuration 🔗

Custom Ports 🔌

🏗️ How It Works

📚 Additional Resources

❓ Frequently Asked Questions

💰 Earning & Payments

🎮 GPU & Hardware

⚙️ Setup & Configuration

🔒 Security & Privacy

🐛 Troubleshooting

🌐 Network & Connectivity

📊 Models & Performance

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages