EVE AI : AI FOR HER HEALTH

This is a submission for the KendoReact Free Components Challenge.

What I Built

EVE / NARI AI is a clinical assistant for women’s health that helps clinicians analyze breast images, prioritize triage, and generate reports, while giving patients clear, patient‑friendly explanations. It includes a RAG-based Breast Cancer Knowledge Box so users can ask grounded questions with context. The app supports role-based workflows (Clinician/Patient), image analysis with BI-RADS assessment, and a dashboard for case review.

KEY FEATURES

AI-Powered Clinical Analysis (for Medical Professionals)

• Breast Scan Analysis: Utilizes AI to highlight suspicious regions, provide BI-RADS scores, generate clinical summaries, and assess risk levels.
• Cephalometric Analysis: Accurately identifies precise anatomical landmarks on X-rays, crucial for orthodontic imaging and planning. Skin Photo Analysis: Identifies lesions, assesses their characteristics (asymmetry, border, color, diameter, evolution), and provides malignancy risk scores.
• Differential Diagnosis Generation: Suggests the most likely differential diagnoses based on provided clinical summaries.
• Longitudinal Summary Generation: Compares current and prior medical scans to summarize changes in findings over time.
• Clinical Note Drafting: Automates the generation of structured clinical notes (e.g., SOAP format) based on AI analysis.
• Second Opinion Generation: Offers AI-powered second opinions on medical scans and initial AI analyses for enhanced diagnostic confidence.
• Clinical and Patient Report Generation: Creates detailed clinical reports for healthcare providers and simplified, patient-friendly reports for better patient understanding.

Patient-Facing AI Tools (for Women Patients )

• Menstrual & Reproductive Health Tracking: Enables tracking of menstrual cycles, prediction of ovulation, and AI-powered insights into symptoms.
• Pregnancy & Postpartum Tracking: Monitors fetal development, common maternal symptoms, and provides tailored health recommendations.
• AI Symptom Checker: Offers preliminary AI analysis and suggests next steps based on reported symptoms.
• Personalized Health Education: Provides clear, illustrated answers to health questions, making complex medical information accessible.
• Medication Helper: Explains prescription information in simple, easy-to-understand terms.
• Scan Annotator: Identifies and labels anatomical structures on medical scans for educational purposes, helping patients understand their own imaging.
• AI Health Plan Coach: Assists users in creating and tracking personalized wellness plans to achieve their health goals.
• Journal Entry Analysis: Summarizes journal entries, identifies overall sentiment, and offers gentle, encouraging suggestions for reflection.

RAG Knowledge Box (Progress Nuclia Integration)

• Interactive Chat Interface: Provides a chat-based interface for users to ask breast cancer-related questions.
• Contextual Answers: Powered by Progress Nuclia, it offers Retrieval-Augmented Generation (RAG) to deliver grounded answers with relevant citations and context.
• Direct Widget Integration: Seamlessly integrated using the official Nuclia widget for robust and reliable knowledge retrieval. Multi-Modal AI Tools
• Image Editing with Text: Allows for editing and annotating images based on natural language text prompts.
• Educational Image Generation: Generates minimalist, medically accurate illustrations for various educational content needs.

TECH STACK

EVE AI is built with a modern and robust technology stack:

• Frontend: React, TypeScript, Vite
• UI Framework: Material UI (MUI) for a consistent and responsive user interface.
• AI Integration: Google Gemini API (@google/genai) for advanced AI capabilities across various features.
• Database: Dexie.js, a powerful IndexedDB wrapper, for efficient client-side data storage.
• UI Components: Leverages a comprehensive suite of KendoReact components from Progress for rich and interactive UI elements, including:

1. PanelBar, PanelBarItem: For collapsible and organized content sections.
2. Scheduler: For appointment management and display.
3. DatePicker: For intuitive date input.
4. Notification: For displaying timely alerts and messages.
5. Chart, ChartSeries, ChartSeriesItem, ChartCategoryAxis, ChartCategoryAxisItem, ChartValueAxis, ChartValueAxisItem: For visualizing health analytics data.
6. Button: For interactive actions.
7. Input: For text input fields.
8. DropDownList: For selection of options from a list.
9. Upload: For handling file uploads.
10. TreeView: For hierarchical navigation in the sidebar.
11. Splitter, SplitterPane: For flexible layout management, especially in analysis views.
12. ProgressBar: For indicating loading and processing status.
13. TabStrip, TabStripTab: For organizing content into tabbed interfaces.
14. Tooltip: For providing additional information on hover.
15. ListView: For displaying lists of items with custom rendering.

• RAG (Retrieval-Augmented Generation): Progress Nuclia, integrated via its official widget, for intelligent knowledge retrieval.

Demo

demo link : https://eve-ai-kappa.vercel.app/

github link : https://github.com/saurkan/eve-ai

KendoReact Components Used

• Buttons, Input
• DropDownList
• Grid
• Upload
• ProgressBar
• TreeView
• Scheduler
• Charts
• Notification
• Tooltip
• Splitter
• TabStrip
• PanelBar
• Menu
• DatePicker
• ListView

AI Coding Assistant Usage

• Used an AI assistant to scaffold the UI, wire KendoReact components, and implement role-based views.
• Automated integration of license activation, theme import, and component swaps from MUI to Kendo to reach 20+ components.
• Assisted in adding a user-facing Gemini API key dialog and updating copy on the landing page.

Nuclia Integration

• Integrated a Breast Cancer Knowledge Box that calls a Nuclia Progress knowledge base endpoint for RAG answers with grounded context.
• The RAG chat works independently of the Gemini key (Gemini is only needed for image analysis/report generation).