Inspire CLI

Command-line interface for the Inspire HPC training platform.

Installation

# Via SSH (recommended)
uv tool install git+ssh://git@github.com/EmbodiedForge/Inspire-cli.git

# Or via HTTPS
uv tool install git+https://github.com/EmbodiedForge/Inspire-cli.git

Local Development

uv tool install -e .
inspire --help

Quick Start

1. Set Credentials

export INSPIRE_USERNAME="your_username"
export INSPIRE_PASSWORD="your_password"
export INSPIRE_TARGET_DIR="/path/to/shared/filesystem"

2. Initialize Config

inspire init

3. Check Resources

inspire resources list

Commands

Command	Description
inspire job create	Submit a training job
inspire job status <id>	Check job status
inspire job logs <id>	View job logs
inspire job list	List cached jobs
inspire notebook list	List notebook instances
inspire notebook create	Create interactive notebook
inspire notebook ssh <id>	SSH into notebook
inspire notebook start <id>	Start stopped notebook
inspire resources list	View GPU availability
inspire run "<cmd>"	Quick job with auto resource selection
inspire sync	Sync code to shared filesystem
inspire bridge exec "<cmd>"	Run command via Git Actions

Examples

# Submit a training job
inspire job create --name "train-v1" --resource "4xH200" --command "bash train.sh"

# Quick run with auto-selected resources
inspire run "python train.py --epochs 100"

# Check GPU availability
inspire resources list

# SSH into a notebook
inspire notebook ssh <notebook-id>

# Sync and run remotely
inspire sync && inspire bridge exec "cd /target && bash train.sh"

Configuration

Config files are loaded from:

• Global: ~/.config/inspire/config.toml
• Project: ./.inspire/config.toml

Download the official config template (internal network only):

curl https://nc.sii.e-forge.org/public.php/dav/files/b42pj6oxfY7ikrM -o config.toml.example

Example config.toml:

[auth]
username = "your_username"

[api]
base_url = "https://your-inspire-platform.com"

[workspaces]
# cpu = "ws-..."       # Default workspace (CPU jobs / notebooks)
# gpu = "ws-..."       # GPU workspace (H100/H200 jobs)
# internet = "ws-..."  # Internet-enabled GPU workspace (e.g. RTX 4090)
# special = "ws-..."   # Custom alias (use with --workspace special)

[[compute_groups]]
name = "H100 Cluster"
id = "lcg-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx"
gpu_type = "H100"

View current config:

inspire config show

Environment Variables

Variable	Description
INSPIRE_USERNAME	Platform username
INSPIRE_PASSWORD	Platform password
INSPIRE_BASE_URL	API base URL
INSPIRE_TARGET_DIR	Shared filesystem path
INSPIRE_WORKSPACE_ID	Default workspace ID
INSPIRE_WORKSPACE_CPU_ID	CPU workspace ID (default workspace)
INSPIRE_WORKSPACE_GPU_ID	GPU workspace ID (H100/H200)
INSPIRE_WORKSPACE_INTERNET_ID	Internet-enabled workspace ID (e.g. RTX 4090)
INSPIRE_PROJECT_ID	Default project ID
INSP_IMAGE	Default Docker image
INSP_PRIORITY	Job priority (1-10)