Description
Reconstructs Hyperliquid’s L4 order book from the raw official node files, down to individual Order ID.
An ETL engine that downloads the raw LZ4 files from Hyperliquid’s official node-data bucket, replays the order-status events, and exports the result to Apache Parquet (ZSTD) — with visibility at the individual-order level, not the aggregated L2 view you get from the UI.
Who it’s for: quants and data teams who need real microstructure, not aggregated L2.
Requirements: Python. Your own AWS account — the official bucket is requester-pays, so transfer costs are billed to whoever downloads. Disk space for the Parquet output.
Status: 4/4 unit tests passing. The full pipeline hasn’t yet been run end-to-end against the live bucket — stated here for transparency.
Licensing: Hyperliquid hasn’t published an explicit redistribution clause for this data. This is a known, shared risk across every provider in the space.
Reviews
There are no reviews yet.