0xArchive comparison · Guide · Free Hyperliquid historical data
Start with 0xArchive Free for Hyperliquid historical data: $0, no AWS account, and the latest 30 days of every route by REST and WebSocket, including core L4 with order IDs and wallets.
- Capacity: 50,000 monthly credits and 15 requests per second.
- Full archive: Build opens the full retained archive for $49 a month.
Free includes live WebSocket feeds and the latest 30 days of REST and replay history, including Hyperliquid L4 and Lighter L3.
How are free Hyperliquid data sources ranked?
- Access effort: an API key and a request, or an AWS account and file parsing.
- Data types included free: book, trades, funding, open interest, liquidations, candles.
- Order-level depth: whether the free tier includes resting orders with wallets.
- History window: the earliest date or the rolling window.
- Support and path to full history: docs, community or email, and what the next tier costs.
0xArchive ranks first because it meets the first three criteria together: key-based access, every data type, and order-level depth on a free plan. Its window is the most recent 30 days; the table gives every source's window.
Which free Hyperliquid historical data sources exist?
| Free source | Hyperliquid data included | History from | How you get it | Order-level book | Limit |
|---|---|---|---|---|---|
| 0xArchive Free | Order book, trades, funding, open interest, liquidations, candles, L4 and order events; Spot, HIP-3, HIP-4; plus Lighter | The latest 30 days on every route (30-day span per request or replay) | REST, live and replay WebSocket, Python, TypeScript and Rust SDKs, CLI, hosted MCP; no AWS account | Yes, L4 with order ID and wallet | 50k credits a month, 15 requests a second, docs support |
| Hyperliquid native S3 | L2 book snapshots, asset contexts, fills and node data | None listed | Requester-pays S3, LZ4 files | No | Uploaded about once a month; "data may be missing"; you pay transfer |
| Hydromancer Reservoir | Fills with liquidations, ADLs, builder and TWAP fills; 1-second candles; daily positions; 20-level book snapshots | Fills 28 Jul 2025; one-minute 20-level books 15 Dec 2025 | Requester-pays S3, no API key, updated daily | None listed | Book snapshots once a minute; requests via Discord |
| CryptoHFTData | Order book, trades, open interest, mark price (futures); spot book and trades | 29 Sep 2025 | Free API key, hourly Parquet files, Python SDK | No (L2 only) | Hourly files; price-level books |
| Tardis.dev free samples | Raw exchange messages as CSV: incremental book, trades, derivative ticker | First day of each month (free sample); trial covers days 1 to 10 of the previous month | CSV and replay API without a key for first-of-month data | None listed | Samples and a 20 GB, one-month trial; no other dates |
| Bitquery free samples | One free sample file per Parquet table (trades, order updates, book updates, funding) | Sample dated 30 Sep 2026 | Parquet files | Per-order book updates (sample) | A sample, not a history; full months cost $600 each |
| GoldRush free trial | HyperCore fills, funding, ledger via Info API replacement | 25 May 2025 | REST and WebSocket | Live l4Book | 14-day trial, 25,000 credits |
Provider facts from each provider's own documentation and pricing, 3 November 2026.
Sources: pricing, venue coverage; Hyperliquid historical data docs; Reservoir docs, archive page, API pricing; Hyperliquid page, order book dataset, Lighter page; Tardis billing FAQ; Bitquery Parquet dataset; GoldRush overview.
What does a free request look like?
A free key reaches every route for the latest 30 days. This BTC trade query covers a five-minute interval on 3 November 2026; the first two rows of the response are shown, each fill carrying the wallet on its side.
curl -G "https://api.0xarchive.io/v1/hyperliquid/trades/BTC" \
-H "X-API-Key: $OXARCHIVE_API_KEY" \
--data-urlencode start=2026-11-03T16:40:56Z \
--data-urlencode end=2026-11-03T16:45:56Z \
--data-urlencode limit=3{
"success": true,
"data": [
{"coin":"BTC","side":"B","price":"77561","size":"0.00016",
"timestamp":"2026-11-03T16:40:56.015Z","direction":"Close Short",
"user_address":"0x36ee...e9f561","fee":"0.005361","fee_token":"USDC"},
{"coin":"BTC","side":"A","price":"77561","size":"0.00016",
"timestamp":"2026-11-03T16:40:56.015Z","direction":"Close Long",
"user_address":"0xbb16...3ad", "closed_pnl":"-0.00224"}
]
}Returned rows draw on the 50,000 monthly credits of the Free plan (a request costs at least one credit), so a three-row request costs one credit.
Trade rows are metered at 1,000 rows per credit; one million rows pulled in 1,000-row pages use 1,000 credits, with rounding per request. The Free plan caps each request or replay at a 30-day span.
What can you do with the Free plan?
- Pull the last 30 days of order books, trades, funding, open interest and liquidations for core perps, Spot, HIP-3 and HIP-4, and trades, books, candles, funding and open interest for Lighter (liquidations from 10 June 2026).
- Rebuild the L4 book at any moment in the window, with order ID and wallet on every resting order.
- Replay a 30-day window over WebSocket at 10x speed, with 10 subscriptions.
- Connect a coding agent through the hosted MCP server, the CLI or the SDKs.
- Test your parser and schema on live routes before you pay.
When does a backtest need more than 30 days?
When your study spans months or years, Build opens each dataset's full retained archive: native order book from 15 April 2023 and trades from April 2023 (every fill from 22 March 2025), funding and open interest from 20 May 2023, liquidations from 22 December 2025 and L4 from 11 March 2026.
Build is $49 a month, or $39 a month billed annually, with 80 million credits, 50 requests a second, 50x replay and $50 in export credits each month. Free and Build use the same routes; the history window, capacity and exports change.
Hydromancer Reservoir vs Hyperliquid S3: how do the free file paths compare?
Hyperliquid's own S3 bucket and Hydromancer's Reservoir are requester-pays file archives: you need an AWS account, you pay transfer, and you parse the files yourself. 0xArchive returns the same markets by symbol and UTC window through one schema.
How do the two free Hyperliquid S3 paths differ?
| You need | Hyperliquid S3 | Hydromancer Reservoir | 0xArchive |
|---|---|---|---|
| Buckets and layout | s3://hyperliquid-archive/market_data/[date]/[hour]/[datatype]/[coin].lz4 and asset_ctxs/[date].csv.lz4; fills and node data in s3://hl-mainnet-node-data/ | s3://hydromancer-reservoir/by_dex/... and global/..., Parquet in date=YYYY-MM-DD partitions | No bucket: REST, WebSocket replay, Parquet exports through the Data Catalog |
| Freshness | "Uploaded ... approximately once a month. There is no guarantee of timely updates and data may be missing." | "The data is updated daily." | Live and historical on one API; status at 0xarchive.io/status |
| Order book | L2 snapshots in market_data; no start date listed | 20-level L2 snapshots at 1-minute cadence, from 15 Dec 2025 (perps, spot, HIP-3 and HIP-4) | Native L2 (up to 20 levels per side) from 15 Apr 2023 by timestamp; full-depth L2 and wallet-attributed L4 from 11 Mar 2026 |
| Trades and fills | Fills in node_fills_by_block; older data in node_fills (API format) and node_trades (different format) | Fills with liquidations, ADLs, builder fills and TWAP fills, from 28 Jul 2025 | Trades from Apr 2023, every fill from 22 Mar 2025, wallet on each row |
| Candles | "No other historical data sets are provided via S3 (e.g. candles or spot asset data)." | 1-second candles from 31 Jul 2025 | Candles 1 minute to 1 week from 22 Mar 2025 |
| Funding and open interest | Asset contexts (asset_ctxs); funding events in misc_events_by_block | None listed | Dedicated routes from 20 May 2023 |
| Liquidations | Inside fill records; no separate dataset listed | Liquidation fills from 28 Jul 2025 | Liquidation events from 22 Dec 2025, with liquidated and liquidator wallets |
| Resting orders with wallets | Not in l2Book; L1 transactions in replica_cmds to decode yourself | None listed | L4: order ID and wallet on every resting order from 11 Mar 2026 |
| Cost | "The requester of the data must pay for transfer costs." | Requester pays; its page says AWS does not charge for the first 100 GB a month sent to the internet | Free $0 for the latest 30 days; Build $49 a month for each dataset's full retained archive |
| Support | "There is no guarantee of timely updates" | Hydromancer Discord for questions and data requests | Docs on Free, email on Build, priority email on Pro and Scale |
The table uses Hydromancer's dataset-specific start dates: fills from 28 July 2025, candles from 31 July 2025 and 20-level one-minute L2 books from 15 December 2025.
Sources: Hyperliquid historical data, Reservoir, Reservoir Hyperliquid, Hydromancer archive page, 0xArchive pricing and venue coverage.
How does requester-pays S3 work for Hyperliquid data?
You need an AWS account and every command carries a flag. Hyperliquid's docs use the AWS CLI and LZ4: copy a file with --request-payer requester, then decompress it. This is the example from its docs, a SOL book snapshot for the 9:00 hour of 16 September 2023:
aws s3 cp s3://hyperliquid-archive/market_data/20230916/9/l2Book/SOL.lz4 /tmp/SOL.lz4 --request-payer requester
unlz4 --rm /tmp/SOL.lz4
head /tmp/SOLHydromancer's bucket uses the same flag and adds Parquet that DuckDB reads directly from S3 (its example sets the region to ap-northeast-1, Tokyo):
aws s3 ls s3://hydromancer-reservoir/by_dex/ --request-payer requesterHydromancer's page gives sizes measured on 18 October 2026: perp fills are about 360 MB a day, 1-second candles about 44 MB a day, liquidation fills about 1 MB a day, and every 1-second candle ever recorded about 15.5 GB.
Those are the quantities that set your transfer bill; the price per GB is AWS's, not Hyperliquid's or Hydromancer's.
What are the gaps in the free Hyperliquid files?
- Hyperliquid's bucket: uploaded approximately once a month, no guarantee of timely updates, data may be missing, and no candles or spot asset data.
- Hydromancer's Reservoir: fills from 28 July 2025, candles from 31 July 2025, the book from 15 December 2025, so a study that starts earlier has no file to read.
- Both: no API, so a missing day is a missing file and you reconcile formats yourself.
0xArchive returns the Hyperliquid native book from 15 April 2023 and trades from April 2023 (every fill from 22 March 2025) through one schema, by symbol and UTC window.
How do you get Hyperliquid data from before the free files begin?
Ask for it by date. This request asks for BTC trades on 20 July 2025, eight days before Reservoir's first fill day. Real request, with the first rows of the response:
curl -G "https://api.0xarchive.io/v1/hyperliquid/trades/BTC" \
-H "X-API-Key: $OXARCHIVE_API_KEY" \
--data-urlencode start=2025-07-20T00:00:00Z \
--data-urlencode end=2025-07-20T01:00:00Z \
--data-urlencode limit=2
{
"success": true,
"data": [
{"coin": "BTC", "side": "B", "price": "117880", "size": "0.02431",
"timestamp": "2025-07-20T00:00:00.806Z", "direction": "Open Long",
"user_address": "0x744c...5f79", "fee": "-0.085969", "fee_token": "USDC"},
{"coin": "BTC", "side": "A", "price": "117880", "size": "0.02431",
"timestamp": "2025-07-20T00:00:00.806Z", "direction": "Open Short",
"user_address": "0xf55d...dee81", "fee": "1.237966", "fee_token": "USDC"}
],
"meta": {"count": 2, "has_more": true, "venue": "hyperliquid"}
}That window is outside the 30 days the Free plan covers, so it needs Build ($49 a month). Inside the last 30 days the same request runs on a free key.
How do you migrate from the free files to an API?
Replace each file read with one request:
| You read today | Request instead |
|---|---|
market_data/[date]/[hour]/l2Book/[coin].lz4 | GET /v1/hyperliquid/orderbook/{coin}?timestamp=... for one moment, or /orderbook/{coin}/history?start=...&end=... for a window |
asset_ctxs/[date].csv.lz4 | GET /v1/hyperliquid/openinterest/{coin} and /funding/{coin} (mark, oracle and mid price on open-interest rows) |
node_fills_by_block or Reservoir fills/perp/all/date=... | GET /v1/hyperliquid/trades/{coin}?start=...&end=..., then follow next_cursor |
Reservoir fills/perp/liquidations/date=... | GET /v1/hyperliquid/liquidations/{coin} (events from 22 Dec 2025, with both wallets) |
Reservoir candles/1s | GET /v1/hyperliquid/candles/{coin}?interval=1m (1 minute to 1 week from 22 Mar 2025), or build any bar size from trades |
Reservoir orderbook/1m | The book at any timestamp, not one snapshot a minute: native L2 from 15 Apr 2023, full-depth L2 and L4 from 11 Mar 2026 |
Whole months as files rather than rows over REST come from the Data Catalog as Parquet.
- Book grain: Hyperliquid order book historical data and the full provider comparison.
- Every provider: for Lighter and every other provider side by side, use the Hyperliquid and Lighter data providers comparison.
- Paid API: for Hydromancer's paid API, see Hydromancer vs 0xArchive.
Start free with the latest 30 days of Hyperliquid and Lighter history, including L4 and L3, by REST and WebSocket. Build opens each dataset's full retained archive for $49 a month. Start free View API plans and export pricing
Frequently asked questions
Is there free Hyperliquid historical data?
Yes. 0xArchive Free provides the latest 30 days of Hyperliquid and Lighter history at $0 by REST and WebSocket, including Hyperliquid L4 and Lighter L3, with 50,000 monthly credits and 15 requests per second. The rolling window stays available while you use the Free plan.
What is the best free Hyperliquid historical data source?
0xArchive Free: $0, no AWS account, every route for the latest 30 days including L4 depth, with REST, WebSocket, replay, SDKs and MCP. Build at $49 a month opens each dataset's full retained archive.
How much Hyperliquid history does the 0xArchive Free plan include?
0xArchive Free includes the latest 30 days of every route by REST and WebSocket, with a 30-day maximum query or replay span, 50,000 monthly credits and 15 requests per second. Build opens the full retained history with 80 million credits, 50 requests per second and $50 in export credits each month.
How do I get Hyperliquid data older than 30 days?
Upgrade to 0xArchive Build at $49 a month ($39 billed annually). It opens each dataset's full retained archive: native order book from 15 April 2023, trades from April 2023 (every fill from 22 March 2025), funding and open interest from 20 May 2023, liquidations from 22 December 2025 and L4 from 11 March 2026.
Is Hyperliquid's S3 historical data complete?
0xArchive serves the market by time range, with the native book from 15 April 2023 and trades from April 2023, so a missing day is a query, not a missing file. Hyperliquid's docs say its bucket is uploaded about once a month, with no guarantee of timely updates and data that may be missing. It also provides no candles or spot asset data.
Can I download Hyperliquid historical data for free as Parquet?
0xArchive sells Parquet exports per market and UTC dates from $0.50 per uncompressed GB, using a separate export-credit balance. The Free plan includes $0 of export credits, so free access is through the API for the latest 30 days; Build includes $50 in export credits each month.