SymptoFlow is a medical assistant chatbot designed to support breast cancer patients throughout their treatment journey. It leverages a vector database containing dozens of research papers collected from the internet to provide accurate and relevant medical insights.

Additionally, both patients and doctors can continuously input medical records, which are timestamped and stored. When interacting with the chatbot, patients benefit not only from the latest research but also from their own medical history. The system distinguishes between records entered by the patient and those entered by the doctor, ensuring that medical recommendations are personalized and informed by professional insights.

The system is designed to be easily adaptable to other diseases. By modifying the system prompt and feeding it different medical literature, the pipeline can be customized for various medical conditions.

This capability makes SymptoFlow a valuable tool for patients in remote or underserved areas, providing them with an AI-powered assistant for medical guidance while enabling doctors to monitor their condition remotely.

SymptoFlow employs a multi-agent system to enhance the chatbot's capabilities:

• One agent is responsible for answering user questions based on vectorized medical knowledge and patient records.
• Another agent generates meaningful titles for conversations, ensuring better organization and referenceability.

This structure allows for efficient question handling and improved chat management, ensuring users can easily navigate their past interactions.

• Seamless Integration: Easily adaptable for different diseases by modifying the system prompt and medical dataset.
• Comprehensive Medical Assistant: Provides patients with science-backed advice based on up-to-date research.
• Patient & Doctor Record System: Stores timestamped medical records from both patients and doctors.
• Doctor-Patient Record Distinction: The chatbot differentiates between records entered by patients and those by doctors to ensure accurate recommendations.
• Remote Patient Support: Helps patients in underserved regions access critical medical information and enables remote monitoring by doctors.
• Multi-Agent Architecture: One agent answers user questions while another generates structured conversation titles.
• Scalability: Can be expanded for different medical fields by updating the vector database and prompt.

• Backend: Python, Flask, LangChain
• Database: MySQL, Vector Database (FAISS)
• Frontend: React, Material UI

• Python 3.9+
• Node.js 16+
• MySQL

git clone https://github.com/tarikalim/SymptoFlow.git
cd SymptoFlow

Please download docker to setup MySQL server if you don't.

Linux: https://docs.docker.com/engine/install/ubuntu/

sudo docker run -d --name sympoto_db \
  -e MYSQL_ROOT_PASSWORD=sympto_root_password \
  -e MYSQL_DATABASE=sympto \
  -e MYSQL_USER=sympto \
  -e MYSQL_PASSWORD=sympto_password \
  -p 3306:3306 \
  mysql:latest

Windows: https://docs.docker.com/desktop/setup/install/windows-install/

docker run -d --name sympoto_db `
  -e MYSQL_ROOT_PASSWORD=sympto_root_password `
  -e MYSQL_DATABASE=sympto `
  -e MYSQL_USER=sympto `
  -e MYSQL_PASSWORD=sympto_password `
  -p 3306:3306 `
  mysql:latest

Download python if you don't: https://www.python.org/downloads/

cd SymptoFlow
python -m venv .venv
source .venv/bin/activate  # For Windows: .venv\Scripts\activate
pip install -r requirements.txt

Create a .env file and define the necessary environment variables that provided in example.env.

Start the backend (from the SymptoFlow directory):

python -m backend.app

Download the node.js if you don't: https://nodejs.org/en/download

cd frontend
npm install
npm start