oid and user_address, so you can rebuild the book level by level and see who stood where in the queue. Order-level L4 depth runs since March 2026 across Hyperliquid core, Spot, and HIP-3, and since May 2026 on HIP-4. Order history and order flow are separate route families.
This is the deepest market-data layer 0xArchive ships, and it is the one that aggregated feeds cannot reconstruct after the fact.
What L4 Returns
When To Use 0xArchive
L4 is the moat. Aggregated L2 tells you the top 20 levels about every 1.2 seconds; L4 tells you every single resting order behind those levels, with the wallet that placed it. That makes queue analysis and exact book reconstruction possible. Join the separate order-history, order-flow, and trade routes when maker-fill attribution, order status, or lifecycle timing is required. 0xArchive carries L4 book depth for Hyperliquid core, Spot, and HIP-3 since March 2026, and HIP-4 since May 2026 as named JSON. Store the request ID when the response exposes it, and check data quality before a backtest trusts the window. 0xArchive is not the right fit when the venue you need order-level depth for is neither Hyperliquid nor Lighter; on Lighter the order-level layer is L3.How To Choose An L4 Vendor
The claim to test is per-order rows, and it collapses into three checks: does each row carry an order ID and a wallet field, does the vendor state when the order-level window starts, and can the same stream replay historically. A feed that summarizes levels, hides attribution, or only runs live is an L2 product with a deeper name, and no later export recovers the orders it never kept.Check This Yourself
One bounded request settles it. Pull a single page of L4 history and read the rows that come back:oid and user_address, inside the same {success, data, meta} envelope as every other route, with a meta.request_id you can quote when checking a window. The route works on every tier, including Free.