Nous Markets
Bloomberg-lite terminal for crypto/stock arbitrage (shelved)
The Problem
Saw a LinkedIn post about arbitrage opportunities across markets. Realized the tooling didn't exist to spot them quickly. Rapid prototype became something bigger.
The Process
Built as a rapid prototype. Developer-first architecture for potential API revenue-the execution of arbitrage opportunities happens too fast for humans, so the real value is in the data pipeline.
The Outcome
Live but shelved. Functional Bloomberg-lite terminal for cross-market arbitrage. Demonstrated the concept; market timing wasn't right for continued investment.
Design Decisions
| Decision | Why | Engineering Tradeoff |
|---|---|---|
| Unified data normalization | Stock and crypto APIs return wildly different formats | Additional abstraction layer for each data source |
| WebSocket-first architecture | Arbitrage requires real-time data, not polling | Connection management complexity |
| Alert thresholds | Traders can't watch every spread-notify on significant opportunities | False positive tuning required |
STAR Summary
| Situation | Saw a LinkedIn post about arbitrage opportunities across crypto and traditional markets. Price differences existed but no unified tooling to spot them quickly. |
| Task | Build a Bloomberg-lite terminal for cross-market arbitrage. Rapid prototype. Fully AI-built. |
| Action | Built unified data normalization layer (stocks and crypto APIs return wildly different formats). WebSocket-first architecture for real-time data. Dashboard for rapid decision-making. Developer-first API for potential automated execution revenue. |
| Result | Live but shelved. Key learning: arbitrage opportunities get eaten too fast for humans to execute-machines win. The real value is the API, not the dashboard. Demonstrated concept; paused for better market timing. |
The Reality Check: Arbitrage opportunities get eaten too fast for humans to execute. Machines win. The real value isn’t the dashboard-it’s the API.
Data Normalization Challenge
| Source | Characteristics | Challenge |
|---|---|---|
| Stocks | REST APIs, market hours, rate limits | Different quote formats per exchange |
| Crypto | WebSocket streams, 24/7 markets | Varying precision and formats |
Nous Markets normalizes everything into a unified format for cross-market comparison.
Real-Time Pipeline
| Step | Action | Purpose |
|---|---|---|
| 1 | WebSocket connections | Maintain live streams |
| 2 | Normalization | Convert to common format |
| 3 | Spread calculation | Compare across venues |
| 4 | Threshold detection | Flag opportunities |
| 5 | Dashboard push | Update UI instantly |
Why It’s Shelved
| Validated | Learning |
|---|---|
| Technical feasibility | ✓ Works |
| Cross-market data | ✓ Normalized |
| Real-time processing | ✓ Fast enough |
| Human execution | ✗ Too slow-machines eat opportunities |
Future Path: The API (data access for automated traders) is the real product. Dashboard was proof of concept.
Gallery
Find market opportunities across crypto and traditional markets