Features

Order matching engine processing thousands of orders per second with sub-millisecond execution

Spot trading with limit, market, and stop orders across multiple trading pairs

Real-time order book, trade history, and candlestick charts streamed via WebSocket

Hot and cold wallet architecture separating operational funds from long-term custody

KYC and AML compliance workflow built into onboarding

Maker-taker fee model with volume-based tier discounts

REST and WebSocket API for algorithmic trading and third-party integrations

Two-factor authentication and withdrawal address whitelisting as standard

Tech stack

Node.js matching engine with an event-sourced order book for deterministic replay and auditability

Kafka as the backbone for order events, trade events, and ledger updates — decoupling the matching engine from settlement, reporting, and notification consumers

Redis for real-time order book state, session management, and rate limiting

Angular trading interface with WebSocket-driven live updates for order book depth, recent trades, and portfolio positions

PostgreSQL for user accounts, transaction history, and audit records

Separate custody service managing hot wallet signing and cold wallet transfer workflows

What CoinRecoil is built around

An exchange has two hard technical problems that every other feature depends on solving correctly. The first is the matching engine — the component that takes a buy order and a sell order and determines whether and at what price they execute. It must be fast, deterministic, and produce an auditable record that both sides of the trade can verify independently.

The second is the event pipeline. A trade is not just a matching event — it triggers a ledger update for two accounts, a fee calculation, a trade history record, a websocket push to all subscribers watching the order book, and potentially a stop-order evaluation. If any of these consumers is slow or fails, it cannot be allowed to block the matching engine or delay the next trade.

CoinRecoil's architecture separates these concerns cleanly. The Node.js matching engine writes trade events to Kafka and returns immediately. Settlement, reporting, notification, and fee calculation are all independent consumers of the same event stream. A slow reporting consumer falling behind does not affect trade execution latency. The event-sourced order book means any point in time can be reconstructed by replaying the event log — which is the auditability property that a regulated exchange needs.

Redis keeps the live order book state in memory so WebSocket clients get real-time depth updates without hitting the database on every tick. The Angular frontend uses a WebSocket connection from page load, so the order book, recent trades, and portfolio positions update live without polling.

Wallet security runs on a strict separation between hot and cold custody. The hot wallet holds only enough balance for immediate withdrawal demand. The cold wallet transfer workflow requires multi-party authorisation and is not connected to the internet-facing infrastructure.

Who it is for

CoinRecoil is the foundation for teams launching a regulated crypto exchange, for projects that need a white-label trading platform for their token, and for communities building a marketplace around a specific digital asset.