for ChatGPT, Claude, Perplexity…
API REFERENCE

Hyperliquid historical data: limits and coverage checks

Pagination is not proof of full history. The current official Info API documentation states at most 2,000 fills in a response and availability of only the latest 10,000 fills. Check the earliest and latest records returned before relying on annual totals.

Key facts
  • Use the trading account address, including subaccounts separately when relevant. An agent address is not a substitute for the account history.
  • startTime and endTime are millisecond timestamps. Preserve the timezone used when grouping records.
  • Keep source identifiers and fee tokens. Orders and executions are different: one order can produce multiple fills.
  • Funding is a separate query. Matching a net total does not certify coverage.

Which endpoint contains which records?

Query typeRecordsKeep
userFillsByTimeExecuted fills available in the queried rangetime, coin, dir, side, px, sz, closedPnl, fee, feeToken, tid, oid, hash
userFundingFunding entriestime and nested delta fields, including coin and signed usdc
userNonFundingLedgerUpdatesOther account ledger movementstime, event type, assets and amounts; a transfer is not automatically a trading gain

Pagination checks that avoid false completeness

  1. Choose an explicit account and range. Keep the original request parameters and response files.
  2. Inspect timestamps in each response rather than assuming ordering from an old experiment.
  3. Follow the current endpoint behavior to page the available window. If using timestamp cursors, handle boundary overlap and duplicate records; multiple fills can share a timestamp.
  4. Keep trade IDs and order IDs. Do not deduplicate by order ID alone.
  5. Verify the earliest timestamp against known activity and check source-retention limits. Empty pages and matching totals cannot establish full history.
Correction to earlier versions: this page previously generalized an older 45,147-fill case reconstruction into a claim that pagination bypasses history limits. That is not a supported current guarantee. Use the documented availability limit and independently verify coverage for your account.

What to do if records are missing

Keep existing exports and raw responses. Check the address, date range and relevant subaccounts. Investigate the currently available archive-data route or ask the data provider about the missing period; bulk archives may require separate processing and payment for storage or egress. Do not fill gaps with invented rows or assume a support archive is guaranteed.

For a tax-data export

The CSV export guide includes free synthetic samples and the scanner’s exact output fields. For checking a discrepancy, use the PnL, fee and funding checklist. The scanner processes available data and does not certify a complete tax return.

Common questions

How much Hyperliquid fill history does the API provide?

The current Info API documentation states at most 2,000 fills per response, with only the latest 10,000 fills available. Check the current official documentation and the timestamps returned for your account. More pages do not guarantee all historical activity.

Does an empty API response mean I never traded in that period?

No. An empty response may mean no returned records for the queried account and range; it does not prove older records never existed. Compare your own saved records, accounts, timestamps and source retention.

Are funding payments part of userFillsByTime?

No. Retrieve funding using userFunding and preserve its signed delta.usdc value separately from fill-level PnL and fees.

Primary sources

Check the linked sources for the rules, integration limits and API definitions discussed on this page. A general source does not settle every classification question or verify every third-party integration behavior.

  1. Hyperliquid docs — Info endpoint Official read-only query definitions, returned fields and documented history limits for fills, funding and ledger activity.

How this page was produced