Vyrora AI

Backend Frontend LLM Deployed on Render Deployed on Vercel License: MIT

A multi-agent AI platform with a FastAPI backend and a React frontend. A lightweight router directs each request to one of four specialized agents, all powered by Groq's LLM API.

Live App: https://vyrora-ai-agent-platform.vercel.app Backend API: https://vyrora-ai-agent-platform.onrender.com

Note: Backend runs on Render's free tier and spins down after inactivity. First request after idle may take 30–50 seconds. This is a platform limitation, not a bug.

Vyrora AI Home Research Agent in action

Vyrora AI demonstrates a practical, production-deployable multi-agent architecture without unnecessary complexity. Instead of a heavyweight orchestration framework, a simple keyword-based supervisor routes requests — appropriate for four independent agents with no shared state between them.

Documenting what was intentionally left out, and why:

• No vector database / RAG pipeline. Document Q&A passes extracted text directly into the prompt. Sufficient for typical document lengths and avoids vector store overhead on free-tier infrastructure where persistent storage isn't reliable.
• No LangGraph or graph-based orchestration. Four independent agents with no shared state don't need it. A keyword router achieves the same result with fewer failure points.
• No authentication or persistent storage. Every request is stateless by design, keeping the project lightweight and appropriate for its current scope.

Backend

• FastAPI
• Groq API (openai/gpt-oss-120b)
• pypdf (PDF text extraction)

Frontend

• React (Vite)
• Axios

Deployment

• Render (backend)
• Vercel (frontend)

vyrora-ai-agent-platform/
├── backend/
│ ├── agents/
│ │ ├── research_agent.py
│ │ ├── document_agent.py
│ │ ├── task_agent.py # planning agent
│ │ ├── presentation_agent.py
│ │ └── supervisor.py # keyword-based router
│ ├── api/
│ │ └── routes.py # all API endpoints
│ ├── services/
│ │ └── groq_service.py # LLM API wrapper
│ ├── main.py # FastAPI app entrypoint
│ ├── requirements.txt
│ └── .env.example
├── frontend/
│ ├── src/
│ │ ├── App.jsx # full UI, tabbed by agent
│ │ └── main.jsx
│ ├── index.html
│ ├── package.json
│ ├── vite.config.js
│ ├── vercel.json
│ └── .env.example
├── render.yaml
└── README.md

• Python 3.10+
• Node.js 18+
• A free Groq API key

cd backend
pip install -r requirements.txt
cp .env.example .env
# Add your GROQ_API_KEY to .env
uvicorn main:app --reload

Runs at http://127.0.0.1:8000 — API docs at /docs.

cd frontend
npm install
cp .env.example .env
# Set VITE_API_URL=http://127.0.0.1:8000
npm run dev

1. Push repo to GitHub
2. New + → Web Service → connect repo
3. Language: Python 3
4. Root Directory: backend
5. Build Command: pip install -r requirements.txt
6. Start Command: uvicorn main:app --host 0.0.0.0 --port $PORT
7. Add env var: GROQ_API_KEY
8. After frontend deploy, add FRONTEND_ORIGIN = your Vercel URL

render.yaml included for Infrastructure-as-Code import.

1. New Project → import repo
2. Root Directory: frontend
3. Framework: Vite (auto-detected)
4. Add env var: VITE_API_URL = your Render backend URL
5. Deploy

• Persistent vector storage for large document Q&A
• User authentication and session history
• Streaming responses for long-form output

MIT

Rose Sharma GitHub · LinkedIn · Portfolio