Package Your Code

Before deploying a code-based service via the Web Interface or API, you need to package your application as a .tar.gz archive containing your code and dependencies.

Choose Your Packaging Method

You can package your code in two ways:

Using ServiceMaker (Recommended)

ServiceMaker is a command-line tool that automates code preparation for deploying services. It processes standard projects and generates deployment-ready artifacts.

What ServiceMaker automates:

• Builds Docker container images using platform runtime base images.
• Installs and packages project dependencies using uv.
• Generates .tar.gz archives ready for upload to Container Manager.
• Creates Dockerfile configurations tailored to your project.
• Sets up virtual environments that match platform base images.
• Enables local Docker image testing before deployment.
• Optionally publishes container images to container registries.
• Processes standard project formats (pyproject.toml, requirements.txt, package.json).

For installation and usage instructions, see the ServiceMaker repository .

Manual Packaging

If you prefer to package your code manually without ServiceMaker, follow the steps below.

1. Create a project structure with your application code and entry point script.
2. Add a dependency configuration file:
   • For Python: Create a pyproject.toml with your dependencies and Python version requirement.
   • For Node.js: Create a package.json with your dependencies and a start script.
3. Use uv for Python projects (recommended):
   • Ensure your pyproject.toml specifies requires-python matching your target runtime (e.g., ">=3.12" for the py12* base images).
   • List all dependencies in the dependencies array.
   • The platform uses uv to install dependencies during containerization.
4. Create the archive:

   tar -czf myservice.tar.gz myproject/
   

5. Test locally (optional but recommended):
   • Install dependencies using uv sync.
   • Run your entry point script to verify it works before uploading.

Example: Python Project

Project structure:

myproject/
├── pyproject.toml
├── main.py
└── config.json

Example pyproject.toml:

[project]
name = "my-service"
version = "1.0.0"
requires-python = ">=3.12"
dependencies = [
    "fastapi>=0.115.0",
    "uvicorn[standard]>=0.32.0",
]

Create the archive:

tar -czf myservice.tar.gz myproject/

Example: Node.js Project

Project structure:

myproject/
├── package.json
├── index.js
└── config.json

Example package.json:

{
  "name": "my-service",
  "version": "1.0.0",
  "main": "index.js",
  "scripts": {
    "start": "node index.js"
  },
  "engines": {
    "node": ">=22"
  },
  "dependencies": {
    "express": "^5.0.0"
  }
}

Create the archive:

tar -czf myservice.tar.gz myproject/

Next Steps

Once you have your .tar.gz package ready, you can deploy it using:

• Deploy via Web Interface
• Deploy via API