Welcome to AyurGenixAI!
AyurGenixAI is an innovative solution designed to bridge the gap between traditional Ayurvedic medicine and modern healthcare. Utilizing advanced AI technologies, including natural language processing (NLP) and machine learning (ML), AyurGenixAI provides personalized Ayurvedic medication recommendations based on classical texts and user inputs. This integration of ancient wisdom with contemporary medical practices aims to enhance holistic health and wellness.
- Finding reliable and personalized Ayurvedic medication advice is challenging in today's fast-paced world.
- Many people seek natural remedies but struggle to get accurate and tailored recommendations for their specific symptoms and conditions.
- AyurGenixAI uses advanced NLP and ML technologies to provide personalized Ayurvedic medication recommendations.
- Users can input symptoms into a user-friendly interface to receive precise advice on Ayurvedic remedies, promoting overall health and wellness.
- The recommendation can include details such as herbs, dietary changes, lifestyle modifications, and other home remedies.
Traditional Ayurvedic consultations require:
- Manual analysis of symptoms
- Matching them with treatment options
- Selecting personalized medications or remedies
This process is time-consuming, inconsistent, and often dependent on practitioner availability.
-
Symptom Intake via NLP
Users enter symptoms in natural language. The NLP module automatically extracts key symptoms (fever, acidity, fatigue, etc.) without human involvement.
➡️ Automation: Patient intake + preliminary diagnosis is automated. -
Dataset Matching with Custom LLM
Symptoms are matched to curated Ayurvedic data (disease → treatment). The LLM processes inputs and generates tailored treatment suggestions (medicines, herbs, diet).
➡️ Automation: Treatment recommendation system similar to a human doctor's initial consultation. -
Backend Workflow (Django)
Once inputs are submitted, data flows through the NLP module, LLM responds, and the user receives results instantly.
➡️ Automation: Entire consultation loop is automated: intake → analysis → recommendation.
In the modern digital era, users often struggle to access personalized and trustworthy Ayurvedic healthcare. AyurGenixAI addresses these challenges by delivering accurate, evidence-based, and individualized recommendations, digitizing Ayurveda using intelligent systems.
- User Input Analysis: Captures and processes user symptoms and conditions.
- Generative AI-Based Recommendations (LLaMA 3.1): Suggests Ayurvedic medications tailored to individual needs.
- Extensive Database: Utilizes a comprehensive database of Ayurvedic treatments and their uses.
- Interactive Interface: Easy-to-use interface for a seamless user experience.
- Human-in-the-Loop (HITL) Validation Workflow – Expert practitioners validate AI-generated recommendations, ensuring clinical reliability, accuracy, and adherence to Ayurvedic principles.
- Secure and Confidential: Ensures user data privacy and security.
Title: AyurGenixAI: Ayurvedic Dataset
Source: Kaggle Dataset – DOI
Entries: 15,160
Diseases Covered: 447
Attributes: 35 holistic parameters
Derived From: Ancient Ayurvedic scriptures, clinical sources, and validated inputs
This structured dataset powers the AI models by providing rich, diverse, and meaningful health insights.
| Component | Description |
|---|---|
| Frontend | Django templates (HTML/CSS), responsive UI |
| Backend | Django REST Framework, PostgreSQL for user and medical data |
| AI Engine | NLP pipeline + LLaMA 3.1 (Generative AI) for context-aware recommendations |
| Database | PostgreSQL |
| Chatbot | LLaMA-based multilingual, dynamic chatbot |
| Deployment | Render via FastAPI |
| Category | Tools & Frameworks |
|---|---|
| Programming | Python |
| Frameworks | Django, Django REST Framework |
| AI/ML | Scikit-learn, Hugging Face Transformers, PyTorch, Ollama (LLaMA 3.1) |
| NLP | NLTK, LLaMA |
| Database | PostgreSQL |
| Frontend | HTML, CSS |
| Deployment | Render |
| Backend Framework | FastAPI |
- Data Processing: Scripts for preprocessing data and preparing it for model training.
- Model Training: Scripts for training NLP and ML models to analyze symptoms and recommend medications.
- Recommendation Engine: Core logic for generating personalized Ayurvedic medication recommendations.
- Anaconda/Miniconda
- Python 3.x
-
Clone the repository
git clone https://github.com/kittu-122/AyurGenixAI.git cd AyurGenixAI -
Create a new conda environment and install dependencies
conda env create -f environment.yml conda activate ayurveda_env
-
Install NLTK data
import nltk nltk.download('punkt') nltk.download('averaged_perceptron_tagger')
Run the text_processing.py script to preprocess the Ayurvedic text files.
python preprocessing/text_processing.pyRun the nlp_analysis.py script to perform NLP analysis and extract relevant information from the preprocessed text.
python preprocessing/nlp_analysis.pyRun the formulation_engine.py script to generate Ayurvedic drug and formulation recommendations based on input symptoms or pharmacological properties.
python recommendation/formulation_engine.py- Age Group, Gender, Occupation and Lifestyle, Cultural Preferences, Family History
- Disease (English, Hindi, Marathi), Hindi Name, Marathi Name, Symptoms, Diagnosis & Tests, Symptom Severity, Duration of Treatment
- Medical History, Current Medications, Risk Factors, Environmental Factors, Sleep Patterns, Stress Levels, Physical Activity Levels
- Doshas, Constitution/Prakriti, Seasonal Variation, Dietary Habits, Herbal Formulations, Herbal/Alternative Remedies
- Diet and Lifestyle Recommendations, Yoga & Physical Therapy, Medical Intervention, Prevention, Prognosis, Complications, Patient Recommendations
- Allergies (Food/Environmental), Cultural Preferences, Formulation
| Model | Accuracy |
|---|---|
| Random Forest | 55% |
| XGBoost | 85% |
| LLaMA 3.1 | 90% |
- System Usability: High SUS scores in end-user testing
- Clinical Accuracy: Validated by Ayurvedic experts
- Human-in-the-Loop Assurance: Ayurvedic practitioners review and refine AI-generated outputs, ensuring reliability, clinical relevance, and alignment with traditional healing principles.
- Hugging Face Transformers – BERT Fine-Tuning
Use: Best practices for fine-tuning BERT on custom Ayurvedic symptom datasets.
-
Ollama Documentation
Use: Run and deploy LLaMA 3.1 models efficiently for healthcare NLP applications. -
LLaMA Model by Meta
Use: Official resources for LLaMA pre-trained models and their customization.
-
Django Documentation
Use: Build robust, secure, and scalable web apps with REST APIs. -
Django REST Framework
Use: Essential for backend API development and database integration.
- PostgreSQL Docs
Use: Reliable, SQL-based storage for structured Ayurvedic data.
-
PyTorch
Use: Ideal for custom NLP model experimentation and LLaMA fine-tuning. -
Hugging Face Model Hub
Use: Access to pre-trained healthcare NLP models and finetuning utilities.
- Render
Use: Full-stack deployment platform supporting web services, static sites, background workers, and APIs. Automatically deploys from Git repositories.
- FastAPI
Use: Modern, high-performance Python framework for building APIs. Offers asynchronous support, automatic docs (Swagger & ReDoc), and is ideal for scalable applications.
If you use the dataset or AyurGenixAI in your work, please cite:
Kirti Rane, Yadnesh Bhanushali, and Nikita Yadav. (2025). AyurGenixAI: Ayurvedic Dataset [Data set]. Kaggle. https://doi.org/10.34740/KAGGLE/DS/7148374
You can access the deployed AyurGenixAI service using the following link:
👉 AyurGenixAI Web App
Contributions to the project are welcome! Feel free to submit pull requests, report issues, or suggest enhancements to improve AyurGenixAI.
Thank you for choosing this project. Hoping that this project proves useful and delivers a seamless experience for your needs!