Matthew
Jan 30 3:07 PM
network_wifi battery_5_bar
folder_open ~/projects/simplifi-ai
auto_awesome

Simplifi.ai

A React Native mobile app that simplifies complex text, emails, and documents using NLP and OCR - making academic and legal materials easy to understand.

React Native Node.js NLP Google Cloud

Simplifi.ai is a mobile application developed with React Native to simplify complex text, emails, and documents, including academic and legal materials. With Simplifi.ai, users can input or upload text, and the app will return a simplified, easy-to-understand version.

Features

  • Text Simplification - Input or paste complex text and receive simplified output
  • Advanced Search - Input and update the area of topic with additional context
  • Read Aloud - Text-to-speech for the simplified document
  • Sentiment Analysis - Gauge if the text has a positive or negative sentiment
  • Document Upload & OCR - Upload PDFs or images, extract text using OCR and simplify it
  • Email Simplification - Upload email text to get a simplified version
  • Document History - Access previously simplified documents for quick reference

Tech Stack

Frontend (React Native)

  • Wink-NLP - Sentiment analysis on submitted documents
  • React-Native-Paper - UI library for user-friendly interface
  • Expo-speech - Text-to-speech functionality
  • React-Native-Markdown-Renderer - Renders documents in markdown

Backend (Node.js + Express)

  • Multer - Upload large files (images, PDFs)
  • tesseract.js - OCR to extract text from images
  • Spellchecker - Fix common OCR spelling mistakes
  • @google/generative-ai - Google’s Gemini 1.5 Flash LLM
  • pdf-parse - Extract text from PDF uploads

Cloud

  • Google Cloud

Future Work

Looking ahead, I plan to move away from hosted LLM services and explore running local models through services such as Ollama. While I currently lack the necessary hardware for local fine-tuning, I’m exploring services like Runpod to handle this process. Fine-tuning would enable me to better tailor the models to user needs as well as experiment with agent frameworks such as bee-agent to enhance the tooling and error handling of LLMs.