NVIDIA/Quantum-Calibration-Agent-Blueprint

NVIDIA/Quantum-Calibration-Agent-Blueprint

Description: This is a reference agent blueprint for AI-powered quantum device calibration. It provides an intelligent agent interface for discovering, executing, and analyzing quantum calibration experiments with support for automated workflows and vision-based analysis.

Language: TypeScript

License: Apache-2.0

Stars: 54

Forks: 16

Open issues: 2

Created: 2026-03-31T17:42:17Z

Pushed: 2026-04-20T06:53:30Z

Default branch: main

Fork: no

Archived: no

README:

Quantum Calibration Agent Blueprint

Part of NVIDIA Ising. This is a reference agent blueprint for AI-powered quantum device calibration. It provides an intelligent agent interface for discovering, executing, and analyzing quantum calibration experiments with support for automated workflows and vision-based analysis.

 *The Web UI provides a chat interface for natural language interaction with the calibration agent.*

 *The CLI provides a terminal-based interface for quantum calibration experiments.*

What is this?

This is a reference agent blueprint for quantum device calibration that combines:

• Intelligent Experiment Discovery: Automatically find and understand available quantum calibration experiments
• AI-Driven Execution: Run experiments through natural language commands or structured workflows
• Visual Analysis: Inspect plots and data using vision language models (VLMs)
• Workflow Automation: Execute complex multi-step calibration sequences with built-in validation
• Smart Data Management: Store and retrieve experiment results with HDF5 and SQLite

The agent supports multiple LLM providers including NVIDIA, Anthropic, and OpenAI.

Quick Start

Prerequisites

• Python 3.11+
• Node.js 18+ and npm 9+ (for Web UI)
• API Key from one of the supported providers:
• NVIDIA API Catalog (default)
• Anthropic
• OpenAI

Installation

# Clone repository
git clone https://github.com/NVIDIA/Quantum-Calibration-Agent-Blueprint.git
cd Quantum-Calibration-Agent-Blueprint

# Set up Python environment
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -e .

# Install UI dependencies
cd ui && npm install && cd ..

# Configure environment
cp .env.example .env
# Edit .env and add your API key (choose one):
# NVIDIA_API_KEY=nvapi-your-key-here
# ANTHROPIC_API_KEY=sk-ant-your-key-here
# OPENAI_API_KEY=sk-your-key-here

Running QCA

Option 1: Full System (Backend + Web UI)

# Terminal 1 - Start Backend
qca serve

# Terminal 2 - Start Web UI
cd ui && npm run dev

Open http://localhost:3000 in your browser.

Option 2: CLI Only

qca

Option 3: Non-Interactive Commands

qca experiments list
qca experiments run t1_measurement
qca workflow list

Key Features

• Interactive TUI: Rich terminal interface for conversational experiment management
• Web UI: Browser-based chat interface with experiment visualization
• CLI Commands: Direct command-line access to all functionality
• Experiment Scripts: Write Python experiments with automatic parameter discovery
• Workflow Engine: JSON-based workflow definitions with state tracking
• VLM Integration: Analyze plots and experimental data visually
• History Tracking: Complete experiment history with SQLite indexing

CLI Commands

Main Commands

| Command | Description | |---------|-------------| | qca | Launch interactive TUI (default) | | qca serve | Start the backend server | | qca -n "prompt" | Run single prompt non-interactively | | qca -r | Resume previous conversation |

Experiment Management

| Command | Description | |---------|-------------| | qca experiments list | List all available experiments | | qca experiments schema | Show experiment parameter schema | | qca experiments run | Execute an experiment |

Workflow Management

| Command | Description | |---------|-------------| | qca workflow list | List all workflows | | qca workflow show | Display workflow definition | | qca workflow status | Check runtime progress |

History and Data

| Command | Description | |---------|-------------| | qca history list | List past experiment executions | | qca history show | Show detailed experiment results | | qca data arrays | List arrays stored in experiment |

Documentation

Full documentation is available in the docs/ directory:

pip install -e ".[docs]"
cd docs && make html
# Open docs/_build/html/index.html

Development

# Install with test dependencies
pip install -e ".[test]"

# Run tests
pytest

# Run with coverage
pytest --cov=core --cov=tools

License

[Apache License 2.0](LICENSE) - Copyright 2026 NVIDIA Corporation