On-Chain OHLCV Pricer

Problem

Propellyr’s tax calculator depended on a third-party pricing service that was expensive, rate-limited, and a single point of failure for the core product.

Solution

Researched AMMs and liquidity pools, then built a Node.js pipeline that fetches on-chain pool data through Infura and web3.js and computes token prices with the AMM formula (x*y=k), mostly from Uniswap pools. Prices came straight from the blockchain, so no external price API was involved at all.

Architecture

A high-throughput Node.js ingestion service pulls liquidity pool state via Infura and web3.js, applies the AMM formula to derive per-block prices, aggregates them into OHLCV candles, and stores everything in ClickHouse on AWS for the tax calculator to query.

Engineering decisions

• Computed prices from pool reserves with the AMM formula instead of paying an external feed
• ClickHouse for candle storage, since the workload is append-heavy time-series data
• Built for real-time delivery, since the tax product priced transactions at execution time

Technology

Node.js · ClickHouse · AWS · web3.js · Infura

Results

• 40% reduction in operational costs versus the external pricing service
• Removed the third-party dependency from the core tax product

What Daniel learned

A price is a derived value, and deriving it from first principles on-chain turned a vendor bill into infrastructure we owned.