VoltageGPU is a Confidential AI Infrastructure platform. We provide hardware-sealed GPU pods (Intel TDX, H100 / H200 / B200) and an OpenAI-compatible inference API for confidential TEE LLMs such as DeepSeek-R1-TEE, Qwen3-32B-TEE and Qwen3-235B-TEE.

Is the VoltageGPU Inference API OpenAI compatible?

Yes. The Inference API is fully OpenAI-compatible. Point any existing OpenAI SDK at https://api.voltagegpu.com/v1 to run prompts against confidential TEE LLMs.

Which models does the Inference API expose?

The Inference API exposes confidential TEE LLMs only: DeepSeek-R1-0528-TEE (Enterprise), Qwen3-235B-TEE (Pro) and Qwen3-32B-TEE (Starter). All inference runs inside Intel TDX enclaves with hardware attestation.

VoltageGPU CLI

Deploy and manage confidential Intel TDX GPU pods from the terminal

Quick Start Features Installation Python SDK

What is VoltageGPU?

VoltageGPU is a Confidential AI Infrastructure platform. Every pod runs inside an Intel TDX hardware enclave with encrypted memory, Protected PCIe and on-chain attestation — data is sealed from the host operator and the hypervisor by design.

The CLI is a thin wrapper over the Volt REST API, giving you one-command access to confidential pod deploys, live inventory, SSH key management and balance tracking.

Quick Start

# Install the CLI
pip install voltagegpu-cli

# Configure your API key
export VOLT_API_KEY="your_api_key_here"

# List available confidential inventory (Intel TDX)
volt cc inventory

# Deploy a confidential pod
volt cc deploy --name my-secure-pod --resource h200-small

# SSH into your pod
ssh <workload-uid>@ssh.deployments.targon.com

Get your API key at voltagegpu.com/dashboard

Features

Confidential Pod Management

Deploy and destroy Intel TDX pods
Live inventory with volt cc inventory
Hardware-attested enclaves by default

Any Docker Image

PyTorch, TensorFlow, JAX, vLLM
Bring your own private registry image
No template required for confidential deploys

SSH Key Management

Add and manage SSH keys
Secure access to your pods

Cost Tracking

Real-time balance monitoring
Per-pod cost breakdown
Usage history

Python SDK

Full programmatic access
Async support
Type hints included

Installation

From PyPI (Recommended)

pip install voltagegpu-cli

From Source

git clone https://github.com/Jabsama/VOLTAGEGPU-CLI.git
cd VOLTAGEGPU-CLI
pip install -e .

Requirements

Python 3.8+
pip

Configuration

Option 1: Environment Variable (Recommended)

export VOLT_API_KEY="your_api_key_here"

Option 2: Configuration File

Create ~/.volt/config.ini:

[api]
api_key = your_api_key_here

Option 3: Pass Directly

from volt import VoltageGPUClient

client = VoltageGPUClient(api_key="your_api_key_here")

Usage

Pods

# List all your pods (confidential pods included)
volt pods list

# Get pod details
volt pods get <pod_id>

# Stop (= destroy enclave) a running confidential pod
volt pods stop <pod_id>

# Delete a pod
volt pods delete <pod_id> --yes

# Get SSH command for a pod
volt pods ssh <pod_id>

To create a pod, use the dedicated volt cc deploy command below — confidential pods are the only pod type exposed publicly.

Templates

Curated Confidential Compute templates shipped by the provider (Jupyter, Ubuntu, DeepSeek-R1, GLM-4.6, Kimi-K2, Qwen3-Coder, GPT-OSS-120B, MiniMax, etc.). Deploy directly with volt cc deploy --template TPL-UID.

# List all curated templates
volt templates list

# Filter by type (RENTAL or SERVERLESS)
volt templates list --type RENTAL

# Machine-readable output
volt templates list --json

# Show the full manifest for one template
volt templates get <tpl-uid>

SSH Keys

# List your SSH keys
volt ssh-keys list

# Add a new SSH key from file
volt ssh-keys add --name "my-laptop" --file ~/.ssh/id_ed25519.pub

# Add a new SSH key directly
volt ssh-keys add --name "my-key" --key "ssh-ed25519 AAAA..."

# Delete an SSH key
volt ssh-keys delete <key_id>

Machines

Lists all Confidential GPU machine tiers with live pricing (Intel TDX hardware only — RTX 4090 and other non-confidential tiers are excluded). Equivalent to volt cc inventory but with a flat table format and ready for scripting.

# List available Confidential machines
volt machines list

# JSON output
volt machines list --json

Confidential Compute

Deploy hardware-attested Intel TDX GPU pods:

# List available Confidential Compute inventory
volt cc inventory

# Option A — deploy with an explicit image + resource
volt cc deploy --name my-secure-pod --resource h200-small --image ubuntu:22.04
volt cc deploy --name training --resource b200-small --image pytorch/pytorch:2.5.1-cuda12.4-cudnn9-devel

# Option B — deploy from a curated template
volt cc deploy --name deepseek-infer --template tpl-5npuuq70m1uo

# Override the template image while keeping the resource + commands
volt cc deploy --name my-pod --template tpl-xxx --image my-registry/custom:latest

# JSON output
volt cc inventory --json

Confidential pods use Docker images directly — no template required. SSH access runs through the SSH gateway: ssh <workload-uid>@ssh.deployments.targon.com.

Confidential pods cannot be paused. volt pods stop destroys the enclave and releases the resource, same as volt pods delete.

Account

# Check your balance
volt account balance

# Get account information
volt account info

JSON Output

All list commands support --json for machine-readable output:

volt pods list --json | jq '.[] | select(.status == "running")'

Python SDK

Basic Usage

from volt import VoltageGPUClient

# Initialize client
client = VoltageGPUClient()

# List pods (confidential pods included)
pods = client.list_pods()
for pod in pods:
    print(f"{pod.name}: {pod.status} ({pod.gpu_type})")

# Destroy a confidential pod (stop == delete for TDX)
client.stop_pod(pod_id)

Available Methods

Method	Description
`list_pods()`	List all your pods (confidential included)
`get_pod(pod_id)`	Get pod details
`stop_pod(pod_id)`	Release a confidential pod (destroys the enclave)
`delete_pod(pod_id)`	Delete a pod
`list_ssh_keys()`	List your SSH keys
`add_ssh_key(name, public_key)`	Add a new SSH key
`delete_ssh_key(key_id)`	Delete an SSH key
`list_templates()`	List curated Confidential Compute templates
`get_template(uid)`	Fetch a single template with its full manifest
`list_machines()`	List Confidential GPU machine tiers with live pricing
`list_confidential_inventory()`	Raw confidential inventory (alias for machines)
`create_confidential_pod(name, …)`	Deploy a pod (accepts `template_uid` or explicit `resource_name` + `image`)
`get_balance()`	Get account balance
`get_account_info()`	Get account information

Under the hood, create_confidential_pod() calls POST /api/volt/pods with { "provider": "confidential", ... }. See the API Reference for the raw REST contract.

Examples

Deploy a confidential inference pod

# Pick a resource from live inventory
volt cc inventory

# Deploy
volt cc deploy \
  --name llm-server \
  --resource h200-small \
  --image ghcr.io/my-org/inference:latest

# SSH in
ssh <workload-uid>@ssh.deployments.targon.com

Batch Operations

# Stop all running pods (destroys the enclaves)
volt pods list --json | jq -r '.[] | select(.status == "RUNNING") | .id' | xargs -I {} volt pods stop {}

Environment Variables

Variable	Description	Default
`VOLT_API_KEY`	Your VoltageGPU API key	-
`VOLT_BASE_URL`	API base URL	https://api.voltagegpu.com/api

Documentation

Contributing

We welcome contributions! Please see our Contributing Guide for details.

# Clone the repository
git clone https://github.com/Jabsama/VOLTAGEGPU-CLI.git
cd VOLTAGEGPU-CLI

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

Links

Websitevoltagegpu.com GitHubgithub.com/Jabsama/VOLTAGEGPU-CLI PyPIpypi.org/project/voltagegpu-cli X / Twitter@VoltageGPU Discorddiscord.gg/6HktnD2GUr

Confidential AI, straight from your terminal

Made by the VoltageGPU Team