Summary

Isaac Automator streamlines the deployment of NVIDIA's Isaac Sim, Isaac Lab, and Isaac Lab Arena onto major public clouds (AWS, GCP, Azure, Alibaba Cloud). It provisions fully configured, GPU-accelerated remote workstations, enabling developers and researchers to rapidly set up environments for robotic application development and testing. The tool offers cost-saving features like instance stop/start and workflow enhancements for data management, aiming to reduce setup time and cloud expenditure.

How It Works

This project utilizes Docker to encapsulate the deployment environment, ensuring consistency across different host systems. It automates the provisioning of cloud infrastructure, including GPU-enabled virtual machines, using declarative configurations managed by tools like Terraform and Ansible (implied by repair/import functionalities). Users specify desired Isaac Sim/Lab versions via Git references, and the system handles the installation and configuration, making complex robotics development platforms accessible remotely.

Quick Start & Requirements

• Primary Install/Run:
  • Install Docker: https://docs.docker.com/engine/install/
  • Build container: ./build (Linux/macOS) or docker build --platform linux/x86_64 -t isaac_automator . (Windows).
  • Enter container: ./run
  • Deploy: ./deploy-aws, ./deploy-gcp, ./deploy-azure, ./deploy-alicloud
• Prerequisites:
  • Docker installed.
  • Cloud provider account (AWS, GCP, Azure, Alibaba Cloud).
  • Cloud credentials configured (e.g., AWS IAM permissions like AmazonEC2FullAccess, gcloud auth login, az login).
  • GPU-accelerated instances are required.
• Setup Time: Deployment typically completes "within minutes."

Highlighted Details

• Multi-Cloud Support: Deploys to AWS, GCP, Azure, and Alibaba Cloud.
• AI Agent Integration: Provides specific instructions and guidance for AI agents to operate the tool.
• Cost Optimization: Supports stop/start functionality for cloud VMs, preserving public IP addresses to save costs.
• Pre-built Images: Accelerates deployment by leveraging pre-baked VM images (./image-aws, ./image-azure, ./image-gcp).
• Version Pinning: Allows specifying exact Git references for Isaac Sim, Isaac Lab, and Isaac Lab Arena versions.
• Workflow Utilities: Includes commands for uploading/downloading data (./upload, ./download), managing deployments (./destroy, ./repair), and custom startup scripts (uploads/autorun.sh).

Maintenance & Community

No specific details regarding maintainers, community channels (e.g., Discord, Slack), or roadmap were present in the provided documentation.

Licensing & Compatibility

The specific open-source license for this project was not detailed in the provided README excerpt. Compatibility for commercial use or closed-source linking would depend on the final license.

Limitations & Caveats

• vGPU instances are explicitly not supported on Alibaba Cloud.
• When importing existing deployments, the original SSH private key (key.pem) cannot be recovered from the cloud; a backed-up copy is necessary for restored SSH access.
• Instance types and supported GPU accelerators vary significantly by cloud provider.