Skip to content
datastore.sh

Cost guide

The cheapest Hyperliquid data providers, from free to paid

Updated November 2026

Short answer

The cheapest Hyperliquid data is Hyperliquid's own: a free Info API and public S3 buckets where you pay only AWS transfer costs. The catch is result caps, monthly uploads, and parsing work. Paid options include Bitquery, Dune, Allium, and datastore.sh, which sells 9 typed Parquet tables covering Jul 2025 to Jul 2026 for $2,000.

Key facts

  • Hyperliquid's Info API is free to call, but userFillsByTime returns at most 2,000 fills per response and only the 10,000 most recent fills for an address.
  • Hyperliquid's S3 buckets are requester-pays, the archive is updated approximately monthly with no guarantee of timely updates, and candles and spot data are not included.
  • Dune's Hyperliquid perpetuals tables require a workspace entitlement, and their fills-based data starts on July 27, 2025.
  • Allium documents that about 1% of Hyperliquid traders, those with more than 10,000 trades at its March 2025 backfill, have only their last 10,000 trades from before that date.
  • datastore.sh sells the Hyperliquid dataset at the same prices as every dataset: $200 for the latest month, $500 for six months, $1,000 for twelve months, and $2,000 for full history.

Hyperliquid is unusual: the exchange publishes a large share of its own history for free. That makes the cheapest option easy to name and harder to use. The free sources have caps, gaps, and native formats, so the real cost moves from the invoice to engineering time.

This guide lists every practical source of Hyperliquid historical data in rough order from free to fully managed, with what each costs to use and what it leaves out. It states datastore.sh prices from our checkout configuration and does not quote other vendors' prices, which change without notice. Vendor details were checked in November 2026.

Hyperliquid data sources, from free to fully managed

01 · Official, free

Hyperliquid Info API and WebSocket

Best for:
Live trading and recent history for a handful of addresses
Pricing model:
Free

Strengths

  • Official source, in real time
  • Market and user endpoints, plus WebSocket subscriptions

Watch out for

  • userFillsByTime returns at most 2,000 fills per response and only the 10,000 most recent fills per address
  • Recording anything beyond the caps is your job
Full comparison →

02 · Official, requester-pays

Hyperliquid S3 archives

Best for:
Teams willing to parse raw archive and node files themselves
Pricing model:
No data charge; the requester pays AWS transfer costs

Strengths

  • L2 book snapshots and asset contexts in hyperliquid-archive
  • Fills, explorer blocks, L1 transactions, and non-trade events in hl-mainnet-node-data

Watch out for

  • Archive updated approximately monthly, with no guarantee and possible missing data
  • No candles or spot data, and files arrive in native formats
Full comparison →

03 · Self-operated

Your own non-validating node

Best for:
Teams that want to record everything going forward, under their control
Pricing model:
Your server, storage, and engineering time

Strengths

  • Hyperliquid documents writing fills from a non-validating node
  • Complete control over what is captured from the day it starts

Watch out for

  • History starts when your node starts recording
  • You operate, monitor, and repair it

04 · APIs and Parquet exports

Bitquery

Best for:
Recent Hyperliquid data alongside many other chains
Pricing model:
Fixed price per Data Store dataset

Strengths

  • Parquet exports with a schema and free sample before purchase
  • The same vendor for APIs and streams

Watch out for

  • The Hyperliquid Data Store listing covers the last 30 days; older ranges are on request
Full comparison →

05 · SQL analytics platform

Dune

Best for:
Exploring Hyperliquid perpetuals in SQL, refreshed hourly
Pricing model:
Requires a workspace entitlement, enabled through the Dune team

Strengths

  • Curated perpetuals tables for trades, positions, market metrics, and oracle prices
  • Hourly updates, with first-party oracle prices from May 2023

Watch out for

  • Fills-based data starts July 27, 2025, and funding and open interest start September 27, 2025
  • Gated access rather than a self-serve purchase
Full comparison →

06 · Enterprise data platform

Allium

Best for:
Institutions that need Hyperliquid next to many other chains, with real-time feeds
Pricing model:
Enterprise agreements

Strengths

  • Data from three in-house nodes running concurrently
  • Enriched trades table including HIP-3 markets, and transfers from December 2023
  • Datashares, Kafka or Pub/Sub streams, and real-time APIs

Watch out for

  • About 1% of traders have only their last 10,000 trades from before March 2025
  • Orders exist only from when Allium began running nodes and cannot be backfilled
Full comparison →

07 · Historical dataset delivery

datastore.sh

Best for:
Backtests and research over fills, funding, and ledger activity
Pricing model:
Fixed price per dataset: $200 for the latest month, $500 for six months, $1,000 for twelve months, and $2,000 for full history

Strengths

  • 9 typed Parquet tables covering Jul 2025 to Jul 2026, including swaps, funding, ledger updates, and validator activity
  • Coverage documented per table, with a free 100-row sample for every table
  • Native Hyperliquid identifiers preserved, so rows trace back to source records

Watch out for

  • Coverage runs Jul 2025 to Jul 2026: it starts later than some enterprise sources and does not yet reach the current month
  • Order book snapshots cover a shorter window than trades and funding
  • Historical only; live trading data comes from Hyperliquid's API

Cost and coverage at a glance

What you pay is only half the comparison. The other half is how far back each source goes and what you must build before the data is usable.

Hyperliquid historical data sources compared by cost and coverage
SourceWhat you payHistoryWhat you build
Info APIFreeCapped at the 10,000 most recent fills per addressRecording and storage
S3 archivesAWS transfer costsAs uploaded, about monthly, possibly incompleteParsing, normalization, gap checks
Own nodeServer and engineeringFrom when your node startsThe whole pipeline
BitqueryFixed price per datasetLast 30 days listed; older on requestLittle; Parquet is delivered
DuneWorkspace entitlementFills from July 27, 2025SQL queries
AlliumEnterprise agreementVaries by table; transfers from December 2023Little; tables are managed
datastore.sh$200 to $2,000 per datasetJul 2025 to Jul 2026Little; Parquet is delivered

What free Hyperliquid data really costs

The official sources cost nothing to access and a great deal to finish. The API caps a wallet's fill history at its 10,000 most recent fills, so a complete record of an active trader cannot be rebuilt from it alone. The S3 archives are uploaded about monthly with no guarantee, arrive in native formats, and leave out candles and spot data.

Turning those sources into analysis tables means recording, parsing, aligning funding with fills by market and time, and proving the result has no gaps. For a one-off look at recent data, that work is small. For a backtest over a year of perpetuals, it is usually the largest cost in the project.

datastore.sh Hyperliquid prices

The Hyperliquid dataset is priced like every other dataset. One purchase includes all 9 tables for the chosen window. Coverage currently ends in Jul 2026, so the latest-month window delivers Jul 2026, not the current calendar month, and individual tables start at different dates within the range.

datastore.sh prices per Hyperliquid dataset, by coverage window
Coverage windowPriceWhat it covers
Latest month$200The most recent month of history
Last 6 months$500The most recent six months of history
Last 12 months$1,000The most recent year of history
Full history$2,000All history listed for this dataset
  • Every table in the dataset: one purchase covers them all
  • Partitioned, versioned Parquet files for the selected window
  • Per-table manifests with schema, lineage, and checksums
  • Corrections republished as new versions for the purchased window

Which option is cheapest for your job

Match the source to the job, and the cheapest option is usually obvious.

  • Live trading or a quick look at recent fills: the free Info API and WebSocket.
  • A one-time study and engineers with spare time: the S3 archives, budgeted for parsing work.
  • Recording everything from today onward: your own non-validating node.
  • Exploring perpetuals metrics in SQL: Dune, if you have the entitlement.
  • Hyperliquid inside a multi-chain enterprise feed: Allium.
  • Fills, funding, and ledger activity from Jul 2025 to Jul 2026 as ready tables, bought this week: datastore.sh.

Frequently asked questions

What is the cheapest source of Hyperliquid historical data?

The cheapest Hyperliquid data is Hyperliquid's own: a free Info API and public S3 buckets where you pay only AWS transfer costs. The catch is result caps, monthly uploads, and parsing work. Paid options include Bitquery, Dune, Allium, and datastore.sh, which sells 9 typed Parquet tables covering Jul 2025 to Jul 2026 for $2,000.

Is Hyperliquid historical data free?

Partly. The Info API is free to call, and the S3 buckets are public but requester-pays, so you pay AWS transfer costs. The API caps fill history at the 10,000 most recent fills per address, and the archive is uploaded about monthly with no completeness guarantee, so free data needs significant work before analysis.

How far back does Hyperliquid historical data go?

It depends on the source. Dune's first-party oracle prices start in May 2023 and its fills-based tables on July 27, 2025. Allium's transfers start in December 2023. The datastore.sh Hyperliquid dataset covers Jul 2025 to Jul 2026, with coverage documented per table. The free API is limited by its per-address fill cap.

Does Dune have Hyperliquid data?

Yes. Dune maintains curated Hyperliquid perpetuals tables for trades, positions, market metrics, order books, and oracle prices, updated hourly. Querying them requires an entitlement on your workspace, enabled by contacting the Dune team. Fills-based data starts on July 27, 2025.

How much does the datastore.sh Hyperliquid dataset cost?

The same as every dataset: $200 for the latest month, $500 for six months, $1,000 for twelve months, and $2,000 for full history. One purchase includes all 9 tables for the chosen window as partitioned Parquet with manifests and checksums, and a free 100-row sample is published for every table.

Sources

Vendor descriptions follow each vendor's own public pages, checked November 2026. Other vendors' prices are not quoted because they change without notice. Product names indicate neither partnership nor endorsement.

Start with the data

Skip the indexing project. Run the query.

Browse documented datasets or send the exact protocol, tables, and historical coverage your team needs.