Menu

Post image 1
Post image 2
1 / 2
0

How I built an Offline-First AI App using LLaMA 3 and React

DEV Community·Amit Mishra·20 days ago
#BYeqd8ZG
#engineering#ai#llm#react#fullscreen#offline
Reading 0:00
15s threshold

Hi everyone, I wanted to share a portfolio project I just finished.. # MedVerify — AI-Powered Medicine Authenticator 🛡️ MedVerify is a production-grade, offline-capable Progressive Web Application (PWA) designed to detect counterfeit medicines in rural and low-connectivity environments across India. It uses an edge-deployed 5-layer AI analysis engine (Visual AI, OCR, Barcode/QR, Pricing Intelligence, and Pharmacy Registry Checks) backed by a massive locally-cached CDSCO database. ![Architecture: Zero-Storage, Edge-First] 🚀 Key Features (Enterprise-Grade) 📡 Hybrid "Offline-Second" Architecture : Gracefully degrades from Cloud AI to a local IndexedDB cache when internet is lost. Online Mode (Full Power) : Uses LLaMA 3, EasyOCR, and Python to extract text from photos and answer complex usage questions. Offline Mode (Backup) : AI image extraction and LLM answering are disabled.…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.

Read More