Templates let you define a computer environment once and launch multiple computers from it. Pre-install dependencies, clone repositories, and configure settings.
Quick Start from orgo import Template, Computer # Create a template template = Template( "data-science" ) template.run( "pip install numpy pandas matplotlib" ) template.build() # Launch computers from it computer = Computer( template = "data-science" )
Creating Templates Basic Template from orgo import Template # Initialize with a name template = Template( "my-template" ) # Add commands to run during build template.run( "apt-get update" ) template.run( "apt-get install -y python3-pip" ) template.run( "pip install requests beautifulsoup4" ) # Build and push to registry template.build()
Chained Syntax All methods return self for chaining:
template = Template( "web-scraper" ) template.run( "apt-get update && apt-get install -y chromium-browser" ) \ .run( "pip install selenium playwright" ) \ .env( "DISPLAY" , ":99" ) \ .build()
Template Methods run(command) Execute shell commands during build:
template.run( "apt-get update" ) template.run( "pip install pandas numpy scikit-learn" ) template.run( "npm install -g typescript" )
copy(src, dest) Copy local files into the template:
template.copy( "./config.json" , "/app/config.json" ) template.copy( "./scripts" , "/app/scripts" )
The context parameter in build() controls the build context directory. Files must be within this directory to be copied.
env(key, value) Set environment variables:
template.env( "PYTHONPATH" , "/app" ) template.env( "NODE_ENV" , "production" )
workdir(path) Set the working directory:
clone(repo_url, dest) Clone a git repository:
template.clone( "https://github.com/user/repo.git" ) template.clone( "https://github.com/user/repo.git" , "/app" )
build(context) Build and publish the template:
# Build with current directory as context template.build() # Build with specific context template.build( "./my-app" )
Using Templates By Name After building a template, reference it by name:
computer = Computer( template = "data-science" )
By Object Pass the template object directly (builds automatically if needed):
template = Template( "ml-env" ) template.run( "pip install torch transformers" ) computer = Computer( template = template) # Builds automatically
In a Workspace Store templates in a specific workspace:
template = Template( "prod-env" , workspace = "production" ) template.run( "pip install gunicorn" ) template.build() # Use in the same workspace computer = Computer( workspace = "production" , template = "prod-env" )
Examples Python Data Science from orgo import Template, Computer template = Template( "python-ds" ) template.run( "apt-get update && apt-get install -y python3-pip" ) \ .run( "pip install numpy pandas matplotlib seaborn scikit-learn jupyter" ) \ .env( "PYTHONUNBUFFERED" , "1" ) \ .build() computer = Computer( template = "python-ds" , ram = 8 ) computer.prompt( "Open Jupyter notebook and create a data analysis" )
Node.js Development template = Template( "nodejs-dev" ) template.run( "curl -fsSL https://deb.nodesource.com/setup_20.x | bash -" ) \ .run( "apt-get install -y nodejs" ) \ .run( "npm install -g yarn typescript ts-node" ) \ .workdir( "/app" ) \ .build() computer = Computer( template = "nodejs-dev" )
Web Scraping template = Template( "scraper" ) template.run( "apt-get update" ) \ .run( "apt-get install -y chromium-browser chromium-chromedriver" ) \ .run( "pip install selenium beautifulsoup4 requests" ) \ .build() computer = Computer( template = "scraper" ) computer.prompt( "Scrape the top headlines from news.ycombinator.com" )
Custom Application template = Template( "my-app" ) template.clone( "https://github.com/myorg/myapp.git" , "/app" ) \ .workdir( "/app" ) \ .run( "pip install -r requirements.txt" ) \ .copy( "./config.prod.json" , "/app/config.json" ) \ .env( "APP_ENV" , "production" ) \ .build( "." ) # Use current directory as build context for COPY computer = Computer( template = "my-app" )
Requirements Building templates requires Docker with BuildKit:
# Install Docker # https://docs.docker.com/get-docker/ # Verify installation docker --version docker buildx version
The SDK automatically configures a remote builder that connects to Orgo’s build infrastructure.
Best Practices
Use descriptive template names
# Good template = Template( "python-ml-pytorch" ) template = Template( "nodejs-18-typescript" ) # Avoid template = Template( "env1" ) template = Template( "test" )
Combine commands to reduce layers
# Good - single layer template.run( "apt-get update && apt-get install -y python3-pip git curl" ) # Less efficient - multiple layers template.run( "apt-get update" ) template.run( "apt-get install -y python3-pip" ) template.run( "apt-get install -y git" ) template.run( "apt-get install -y curl" )
Clean up after installing
template.run( "apt-get update && apt-get install -y python3-pip && rm -rf /var/lib/apt/lists/*" )
Pin versions for reproducibility
template.run( "pip install numpy==1.24.0 pandas==2.0.0" )
Troubleshooting
Build fails with 'docker not found'
Install Docker Desktop or Docker Engine:
Build fails with connection error
The build uses Orgo’s remote builder. Ensure you have:
A valid ORGO_API_KEY set
Network access to orgoforge.fly.dev:1234
Template not found when launching computer
Make sure the template was built successfully: template = Template( "my-template" ) template.run( "pip install requests" ) image_ref = template.build() # Check this succeeds print ( f "Built: { image_ref } " ) # Then use it computer = Computer( template = "my-template" )
