ChemCLI - Natural Language Computational Chemistry

A beautiful conversational CLI agent that lets you perform quantum chemistry calculations using natural language, just like ChatGPT or Claude but for computational chemistry.

 ██████╗██╗  ██╗███████╗███╗   ███╗     ██████╗██╗     ██╗
██╔════╝██║  ██║██╔════╝████╗ ████║    ██╔════╝██║     ██║
██║     ███████║█████╗  ██╔████╔██║    ██║     ██║     ██║
██║     ██╔══██║██╔══╝  ██║╚██╔╝██║    ██║     ██║     ██║
╚██████╗██║  ██║███████╗██║ ╚═╝ ██║    ╚██████╗███████╗██║
 ╚═════╝╚═╝  ╚═╝╚══════╝╚═╝     ╚═╝     ╚═════╝╚══════╝╚═╝

Features

🗣️ Natural Language Interface - "Calculate the HOMO-LUMO gap of benzene using DFT"
🎨 Beautiful Terminal UI - Gorgeous ASCII art, colors, and interactive elements
🧪 Auto-Install Chemistry Software - Automatically downloads and installs xTB, PySCF, ORCA
⚡ Multiple Theory Levels - From fast semi-empirical (xTB) to accurate DFT (PySCF/ORCA)
📊 Rich Results Display - Tables, spectra, and molecular properties in your terminal
💾 Results Management - Automatic saving and organization of calculation results
🤖 Multi-LLM Support - Works with OpenAI, Anthropic, or local Ollama models

Quick Start

# Install ChemCLI
npm install -g chem-cli

# Set up your LLM (choose one)
export OPENAI_API_KEY="your-key-here"
export ANTHROPIC_API_KEY="your-key-here"
# or install Ollama for local models

# Start chatting with chemistry
chem-cli

You: Calculate the HOMO-LUMO gap of caffeine using B3LYP

🤖 I'll calculate the HOMO-LUMO gap of caffeine using B3LYP/def2-SVP in PySCF. First let me install PySCF if needed...

[Installing PySCF via pip...]
[Generating 3D structure for caffeine...]
[Running B3LYP/def2-SVP calculation...]

Results:
┌─────────────┬─────────────┐
│ Property    │ Value       │
├─────────────┼─────────────┤
│ HOMO        │ -6.42 eV    │
│ LUMO        │ -0.89 eV    │
│ Gap         │ 5.53 eV     │
│ Method      │ B3LYP       │
│ Basis       │ def2-SVP    │
└─────────────┴─────────────┘

Results saved to: ./results/caffeine_b3lyp_2024-01-15.json

Beautiful Terminal UI

ChemCLI features a gorgeous, professional interface designed for the modern terminal:

🎨 Visual Elements

• Stunning ASCII Art - Beautiful logos and molecular structures
• Color-Coded Results - Energy values, frequencies, and properties with intuitive colors
• Professional Tables - Clean, formatted data presentation with borders
• Interactive Progress - Real-time spinners and progress bars
• Smart Auto-completion - Tab completion for molecules, methods, and commands

🔥 Enhanced Experience

╭──────────────────────────────────────────────────╮
│               🧪 CALCULATION RESULTS              │
├──────────────────────────────────────────────────┤
│ HOMO Energy    │ -6.42 eV     │ Frontier orbital │
│ LUMO Energy    │ -0.89 eV     │ Frontier orbital │
│ HOMO-LUMO Gap  │  5.53 eV     │ Electronic gap   │
│ Dipole Moment  │  0.00 D      │ Molecular moment │
╰──────────────────────────────────────────────────╯

🚀 Status Indicators

⚙️ Chemistry Software Status:
  ✅ ⚡ XTB v6.6.1       - Fast semi-empirical
  ✅ 🐍 PYSCF v2.3.0     - Accurate DFT  
  ❌ 🦋 ORCA             - Professional QC

What You Can Ask

Geometry & Properties

• "Optimize the geometry of water using xTB"
• "What's the dipole moment of acetone?"
• "Calculate molecular orbitals for benzene"

Spectroscopy

• "Show me the IR spectrum of methanol"
• "Calculate UV-Vis absorption of rhodamine B"
• "Predict 1H NMR chemical shifts for glucose"

Thermochemistry

• "What's the binding energy of water dimer?"
• "Calculate reaction energy for methane combustion"
• "Compare stability of chair vs boat cyclohexane"

Method Comparison

• "Run both xTB and B3LYP on ethanol and compare"
• "Which is more accurate for organic molecules: M06-2X or B3LYP?"

Supported Software

Installation

Prerequisites

• Node.js 18+
• Python 3.8+ (for PySCF)
• LLM Access: OpenAI API, Anthropic API, or local Ollama

Install ChemCLI

# Global installation
npm install -g chem-cli

# Or clone and build
git clone https://github.com/your-username/chem-cli
cd chem-cli
npm install
npm run install-global

Set Up LLM Backend

Choose one option:

Option 1: OpenAI (Recommended)

export OPENAI_API_KEY="sk-your-key-here"

Option 2: Anthropic Claude

export ANTHROPIC_API_KEY="sk-ant-your-key-here"

Option 3: Local Models (Free)

# Install Ollama
curl -fsSL https://ollama.ai/install.sh | sh
ollama pull llama2  # or llama3.1

Chemistry Software Setup

ChemCLI will automatically install chemistry software as needed, but you can pre-install:

# xTB (automatic on first use)
# PySCF (automatic on first use)

# ORCA (manual)
# 1. Register at https://orcaforum.kofo.mpg.de/
# 2. Download and extract ORCA
# 3. Add to PATH

Usage Examples

Basic Calculations

chem-cli
> Calculate the energy of water

> Optimize benzene geometry with B3LYP

> What's the HOMO-LUMO gap of caffeine?

Advanced Workflows

> First optimize water with xTB, then run single point B3LYP/def2-TZVP

> Calculate IR spectrum of acetone at B3LYP level with SMD water solvation

> Compare the UV-Vis spectra of anthracene and tetracene using TDDFT

Batch Processing

> Calculate HOMO-LUMO gaps for all molecules in this list: benzene, toluene, xylene

> Optimize geometries of glucose conformers and compare energies

Configuration

Create ~/.chemcli/config.json:

{
  "llm": {
    "provider": "openai",
    "model": "gpt-4"
  },
  "chemistry": {
    "default_software": "xtb",
    "default_method": "gfn2-xtb",
    "default_basis": "def2-svp"
  },
  "output": {
    "save_results": true,
    "results_dir": "./results",
    "format": ["json", "xyz"]
  }
}

File Structure

chem-cli/
├── results/           # Calculation results
├── scratch/           # Temporary files
├── env/              # Auto-installed software
└── config/           # Configuration files

Development

git clone https://github.com/your-username/chem-cli
cd chem-cli
npm install

# Development mode
npm run dev

# Run tests
npm test

# Build for production
npm run build

Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

License

MIT License - see LICENSE for details.

Roadmap

• Paper Reproduction - "Recalculate Table 2 from DOI:10.1021/..."
• Reaction Networks - Multi-step reaction calculations
• Machine Learning - Integration with ML potentials
• Cloud Computing - Integration with HPC/cloud resources
• Jupyter Integration - Export to Jupyter notebooks
• 3D Visualization - Molecular structure visualization

Acknowledgments

• xTB - Fast semi-empirical calculations
• PySCF - Python quantum chemistry
• ORCA - Professional quantum chemistry
• RDKit - Molecular structure generation

ChemCLI - Making computational chemistry accessible through natural language 🧪✨