Research Assistant AI

Research Assistant AI is an advanced, autonomous research platform powered by LangGraph, FastAPI, and React. It utilizes a swarm of specialized AI agents to automate academic research—from planning and searching across academic databases to synthesizing literature reviews and identifying research gaps.

Features

• Multi-Agent Orchestration: Powered by LangGraph, featuring 9 specialized agents (Planner, Search, PDF Processor, Hybrid RAG, Analysis, Citation, Literature Review, Research Gap, and Memory).
• Hybrid RAG: Combines vector search (ChromaDB) with keyword search for highly accurate context retrieval.
• Academic Integrations: Automatically searches and fetches papers from arXiv, Crossref, and Semantic Scholar.
• Real-Time Streaming: Live Server-Sent Events (SSE) streaming of agent thought processes and research generation to the frontend.
• Local LLM Support: Designed to work with Ollama for running models locally and securely.
• Modern UI: A stunning, responsive React frontend built with Vite, Tailwind CSS, Framer Motion, and Zustand.
• Robust Backend: Asynchronous FastAPI backend with PostgreSQL (via asyncpg and SQLAlchemy) for session and paper management.

Tech Stack

Frontend

• React 18 & Vite
• Tailwind CSS & Framer Motion
• Zustand (State Management)
• React Router
• Lucide Icons

Backend

• FastAPI (Python 3.13)
• LangGraph & LangChain
• PostgreSQL & asyncpg (Database)
• Alembic (Migrations)
• ChromaDB (Vector Store)
• Ollama (LLM Provider)
• bcrypt & JWT (Authentication)

Getting Started

Prerequisites

• Python 3.13
• Node.js (v18+)
• PostgreSQL
• Ollama (running locally)

1. Backend Setup

cd backend
# Create virtual environment
python3 -m venv venv
source venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Set up environment variables
cp .env.example .env
# Edit .env with your PostgreSQL credentials

# Run database migrations
alembic upgrade head

# Start the backend server
uvicorn app.main:app --reload --port 8000

2. Frontend Setup

cd frontend
# Install dependencies
npm install

# Start the frontend server
npm run dev

3. Usage

• Open http://localhost:5174/ (or the port specified by Vite).
• Create an account and log in.
• Navigate to the Research Workspace.
• Enter a topic (e.g., "Literature review on federated learning for IoT security") and watch the agents work!

Architecture

1. User Input: A user submits a query via the React frontend.
2. Planner Agent: Breaks down the query into actionable research steps.
3. Search Agent: Queries arXiv, CrossRef, etc., to find relevant papers.
4. PDF Processor: Downloads and chunks PDFs for embedding.
5. Hybrid RAG: Embeds chunks using ChromaDB and retrieves the most relevant context.
6. Analysis & Citation Agents: Analyzes the findings and ensures accurate citations.
7. Synthesis Agents: Generates a comprehensive literature review and identifies gaps in current research.
8. Memory Agent: Persists the conversation and findings to PostgreSQL.

License

MIT License