🧠 Project X

██████╗ ██████╗ ██████╗ ██╗ ██╗ ██╗ ██╔══██╗██╔══██╗██╔═══██╗ ██║ ╚██╗██╔╝ ██████╔╝██████╔╝██║ ██║ ██║ ╚███╔╝ ██╔═══╝ ██╔═══╝ ██║ ██║ ██║ ██╔██╗ ██║ ██║ ╚██████╔╝ ███████╗██╔╝ ██╗ ╚═╝ ╚═╝ ╚═════╝ ╚══════╝╚═╝ ╚═╝ P R O J E C T X

A lightweight AI-powered tool to help researchers prescreen abstracts for systematic reviews. Built with:

• ⚡ FastAPI for the backend
• ⚛️ React for the frontend
• 🧠 Ollama for local LLM inference (e.g., LLaMA 3, Mistral)
• 🗃️ Future-ready for provider APIs and AI agents (MCP, OpenAI, Claude, etc.)

---

📦 Features

• Upload abstracts in bulk (CSV/JSON)
• AI-driven relevance classification
• Confidence scoring for each abstract
• Reviewer UI with filters and label override
• Role-based access control (auth included)
• Future roadmap for full AI agent integration

---

🚀 Quickstart

Backend (FastAPI + Ollama)

cd backend
python -m venv venv
source venv/bin/activate  # or .\venv\Scripts\activate on Windows
pip install -r requirements.txt
uvicorn app.main:app --reload

Make sure Ollama is installed and a model is running:

ollama run mistral  # or any supported model

Frontend (React)

cd frontend
npm install
npm run dev