Ivry - Bloomberg Terminal for Prediction Markets

A Bloomberg Terminal-style platform for prediction markets, creating financial derivatives on top of existing prediction market contracts from Kalshi, Polymarket, and Manifold.

Vision

Ivry creates synthetic financial assets from prediction markets:

• ** Index Products**: Combined baskets of related contracts (e.g., Election Index, Crypto Index)
• ** Volatility Options**: Hedge against prediction market volatility
• ** Single Contracts**: Individual prediction market positions
• ** Basket Products**: Diversified portfolios across themes

Users can search, discover, and trade these financial instruments through a professional trading terminal interface.

Monorepo Structure

arbitrader-1/
├── src/arbitrader/      # Python backend
│   ├── clients/         # API clients for Kalshi, Polymarket, Manifold
│   ├── flows/           # Prefect workflows for data collection
│   ├── models/          # Data models
│   ├── services/        # Embeddings, Qdrant (vector DB), Clustering
│   └── api/             # FastAPI layer (coming soon)
├── frontend/            # Next.js Terminal UI (Dark Mode)
│   ├── app/             # Pages and routes
│   ├── components/      # Terminal-style React components
│   └── lib/             # Types and data models
└── README.md

Setup

1. Install dependencies:

pip install -e .

2. Set up environment variables:

cp .env.example .env
# Edit .env with your API keys if needed

For Qdrant Cloud:

• Set QDRANT_URL to your cluster URL (e.g., https://xxx.qdrant.io)
• Set QDRANT_API_KEY to your API key
• Leave QDRANT_HOST and QDRANT_PORT empty

For local Qdrant:

• Set QDRANT_HOST=localhost and QDRANT_PORT=6333
• Leave QDRANT_URL and QDRANT_API_KEY empty

3. Start Qdrant (if using local instance):

docker run -p 6333:6333 qdrant/qdrant

Running the Flow

Running directly:

python run_flow.py

Using Prefect CLI (Prefect 3.x):

# Deploy the flow
prefect deploy src/arbitrader/flows/market_events.py:process_market_events --name market-events

# Run the deployment
prefect deployment run process_market_events/market-events

Using Prefect Flow Server (for development):

python deploy.py

This will start a flow server that you can trigger via the Prefect UI or API.

Backend Project Structure

• src/arbitrader/ - Main Python package
  • clients/ - API clients for Kalshi, Polymarket, and Manifold
  • flows/ - Prefect workflow definitions
  • models/ - Data models
  • services/ - Embedding, Qdrant, and Clustering services
  • config.py - Configuration settings

Frontend

The Next.js frontend is in the frontend/ directory. See frontend/README.md for details.

cd frontend
npm install
npm run dev  # Runs on http://localhost:3000

Features

• Fetches events from Kalshi and Manifold APIs
• Generates embeddings using MiniLM-L6-V2 model
• Stores events and embeddings in Qdrant vector database
• Uses Prefect for workflow orchestration with retries and error handling

Manifold API

Manifold API works without authentication for public data. Optional API key can be added to .env if needed.

---

SETUP:

BACKEND: python -m venv .venv .venv\Scripts\activate pip install uv pip install google-genai uv sync

proceed with post setup

FRONTEND: cd frontend npm install

POST SETUP

RUN BACKEND: .venv\Scripts\activate uvicorn src.arbitrader.api.server:app --reload --host 0.0.0.0 --port 8000

RUN FRONTEND: cd frontend npm run dev