Piton Studios
Piton Studios
[45]

Holly Trader

2025
[ VIDEO · PROJECT ]
Client
Holly Trader
Year
2025
Scope
Fintech tool, analytics, prototype
Role
Product / Development
§ Case Study

Holly Trader — Trading and analysis tool prototype built on Node.js.

Holly Trader is a Node.js-based tool prototype that monitors market data and produces trading analyses. Data collection, signal generation and reporting flow through a single service.

This fintech product experiment is a side project demonstrating real-time data processing and automation capability.

Node.jsTradingAnalyticsAutomation

Challenge

Holly Trader is a fintech product experiment Piton Studios launched on its own — there is no real exchange client or live capital behind it. The goal was to build, inside a monorepo with clean backend/frontend/package boundaries, an end-to-end trading infrastructure that enforces leveraged-trading risk on the server side, brings multiple asset classes (crypto, equities, commodities) under one roof, and can run automated strategies against real-time market data. The real challenge was keeping market data streamed over WebSocket, order/position state and risk limits in a single, consistently synchronized state.

Solution

The backend was built on NestJS with REST, WebSocket and BullMQ-based queues; Prisma and PostgreSQL handle the data layer, Redis handles queuing and caching. The Binance integration groups spot, futures, wallet and user-stream services into separate modules, with separate market-data services covering equities, commodities and DEX data. Authentication supports classic JWT login alongside ethers-based wallet (MetaMask) signature login. A RiskService validates leverage, exposure and margin ratio server-side before every order; a dedicated stop-loss/take-profit monitor and trailing-stop service check open positions periodically, and a reconciliation service compares exchange state against local data. The strategy engine runs ready-made strategies — grid bot, DCA, RSI mean-reversion, MACD crossover, breakout and scalping — through a shared interface, backed by a separate BullMQ-queued backtesting service. The frontend is built on Next.js 14; live data arriving via socket.io-client feeds into Zustand and React Query state, with lightweight-charts and recharts powering the analysis interface.

Key technical decisions

  • RiskService validates every order server-side against leverage, margin and exposure limits — the check is never left to the client.
  • The Binance integration groups spot, futures, wallet and WebSocket streaming into separate services, including funding and sync jobs.
  • BullMQ- and Redis-based queues process order execution, price alerts and analytics calculations in the background.
  • Eight ready-made strategies (grid bot, DCA, RSI mean-reversion, MACD crossover, breakout, scalping, multi-timeframe, custom) run through a shared strategy interface, backed by a dedicated backtesting engine.
  • Authentication works via JWT or an ethers-based wallet signature (MetaMask); a reconciliation service continuously compares exchange state against local data.

Tech stack

TypeScriptNestJSNext.jsPrismaPostgreSQLRedisBullMQSocket.IOethers.jsTurborepo

Outcome

The result is a trading-infrastructure prototype that unifies crypto, equity and commodity markets in one interface while enforcing risk on the server side. Backend and frontend live in separate packages that share common types through a shared package. The project is a side venture demonstrating our technical capability in real-time data processing, queued job execution and multi-exchange integration — it makes no claim of live capital or trading performance.

Related solutions