AI Research Assistant

A full-stack web app that turns a plain-language question into a synthesized, cited research explanation — pulling live context from Wikipedia and arXiv, then using OpenAI to weave both into a single coherent answer with linked sources.

Live demo: ai-research-assistant-xi-jade.vercel.app (Backend runs on Render's free tier, which spins down after inactivity — the first request after idle time may take 30-60 seconds to respond. Subsequent requests are fast.)

How it works

1. You ask a question in plain language.
2. The backend resolves your query against Wikipedia's search API to find the best-matching article, and separately queries arXiv for relevant academic papers.
3. Both sources are passed to OpenAI (gpt-4.1-mini), which synthesizes them into a single explanation.
4. The answer, along with clickable source citations (tagged by type), is returned and saved to your research history.

Tech stack

Backend: Python, FastAPI, SQLAlchemy, SQLite, pytest Frontend: React, TypeScript, Vite External APIs: OpenAI, Wikipedia, arXiv Deployment: Render (backend), Vercel (frontend)

The backend follows a layered architecture (api → service → provider → repository → db) — see docs/architecture.md for details, and docs/api.md for the endpoint reference.

Features

• Natural-language research queries synthesized from Wikipedia + arXiv
• Source citations, distinguished by type (Wikipedia vs. arXiv)
• Persistent research history with favorite and delete
• Automatic database initialization on startup
• Backend test suite (pytest) covering core endpoints

Running it locally

Backend

python -m venv venv
venv\Scripts\Activate.ps1        # Windows PowerShell
# source venv/bin/activate       # macOS/Linux

pip install -r backend/requirements.txt

Create a .env file in the project root:

APP_NAME=AI Research Assistant ENVIRONMENT=development OPENAI_API_KEY=your-key-here

Then run:

uvicorn backend.app.main:app --reload

The API will be live at http://127.0.0.1:8000 (tables are created automatically on first run).

Frontend

cd frontend
npm install
cp .env.example .env
npm run dev

Open the URL it prints (usually http://localhost:5173).

Tests

pytest

Roadmap

This project is an ongoing learning exercise. Planned next steps:

• Docker support for one-command local setup
• Migrating from SQLite to a persistent hosted database
• User accounts and saved paper collections
• Expanded source integrations (e.g. GitHub repositories)

See docs/architecture.md for the fuller long-term vision.