Quick Start Guide

Get OHMind up and running in under 10 minutes

This guide provides the fastest path to a working OHMind installation. For detailed installation instructions and troubleshooting, see the full installation guide.

Prerequisites
Installation
Verification
Hello World Example
Next Steps

Prerequisites

Before you begin, ensure you have:

Requirement	Version	Check Command
Linux	Ubuntu-like	`uname -a`
Conda	Anaconda/Miniconda	`conda --version`
Python	3.10+	`python --version`
GPU + CUDA	Recommended	`nvidia-smi`

Quick Prerequisite Check

# Run these commands to verify your system is ready
conda --version          # Should show conda version
python --version         # Should show Python 3.10+
nvidia-smi              # Optional: shows GPU info if available

Installation

Step 1: Clone and Navigate

# Clone the repository (if not already done)
git clone <repository-url> OHMind
cd OHMind

Step 2: Create Conda Environment

# Create the OHMind environment from environment.yml
conda env create -f environment.yml

# Activate the environment
conda activate OHMind

This installs all Python dependencies including:

LangChain + LangGraph for agent orchestration
PyTorch, DGL for VAE models
RDKit for cheminformatics
FastAPI, Chainlit for web interfaces
Textual for TUI
langchain-mcp-adapters for MCP integration

Step 3: Configure Environment Variables

Create or edit the .env file in the project root:

# Minimal configuration
OHMind_workspace=/path/to/your/workspace

# LLM Configuration (OpenAI-compatible)
OPENAI_COMPATIBLE_API_KEY=your-api-key
OPENAI_COMPATIBLE_BASE_URL=https://api.openai.com/v1
OPENAI_COMPATIBLE_MODEL=gpt-4

Step 4: Create Workspace Directories

# Create the workspace structure
mkdir -p $OHMind_workspace/{HEM,QM,MD,Multiwfn}

Verification

Verify Installation

Run this command to verify your installation:

# Activate environment and check imports
conda activate OHMind
python -c "
from OHMind.OHVAE import JTPropVAE
from OHMind.OHPSO import BasePSOptimizer
from OHMind.OHScore import metrics
print('✓ OHMind core library loaded successfully')
"

Expected output:

✓ OHMind core library loaded successfully

Verify MCP Servers

Test that MCP servers can start:

# Test Chem server (should show available tools)
python -m OHMind_agent.MCP.Chem.server --help

Hello World Example

Let’s run a simple example to verify everything works.

Option A: Using the CLI (Recommended for Quick Test)

# Start the CLI application
./start_OHMind_cli.sh

Once the CLI starts, try this prompt:

What is the molecular weight of aspirin (acetylsalicylic acid)?

The system should:

Convert the name to SMILES
Calculate the molecular weight
Return the result (~180.16 g/mol)

Option B: Using the Web UI

# Start the backend and UI
./start_apps.sh

Then open http://localhost:8000 in your browser and log in with:

Username: admin
Password: admin

Try the same prompt in the chat interface.

Option C: Using Python Directly

from OHMind.OHVAE import JTPropVAE
from OHMind.OHPSO.util import smiles_to_mol
from rdkit.Chem import Descriptors

# Simple molecular weight calculation
smiles = "CC(=O)OC1=CC=CC=C1C(=O)O"  # Aspirin
mol = smiles_to_mol(smiles)
mw = Descriptors.MolWt(mol)
print(f"Molecular weight of aspirin: {mw:.2f} g/mol")

Expected output:

Molecular weight of aspirin: 180.16 g/mol

Quick Checklist

Use this checklist to ensure your installation is complete:

Conda environment created: conda env create -f environment.yml
Environment activated: conda activate OHMind
Workspace directory exists and is writable
.env file configured with LLM API keys
Verification script runs without errors

Optional External Tools

For full functionality, you may also need:

ORCA installed and OHMind_ORCA environment variable set
Multiwfn installed and MULTIWFN_PATH set
GROMACS available on PATH
Qdrant running for RAG functionality

See the full installation guide for details on setting up external tools.

Next Steps

Now that OHMind is running, explore these resources:

Goal	Resource
Learn the interface	First Steps
Understand the architecture	Architecture Overview
Run HEM optimization	HEM Optimization Tutorial
Configure MCP servers	MCP Configuration
Troubleshoot issues	Troubleshooting Guide