Data pricing
Choose what you want to do. Each tab shows what you get, what it costs and how to start.
The complete archive: choose how long
12 months
$1,290one-off$108 per month of new files- The full archive to date on day one
- New daily files for 12 months
- Downloaded files stay licensed after the term
24 months
Save 15%$2,190one-off$91 per month of new files- The full archive to date on day one
- New daily files for 24 months
- Downloaded files stay licensed after the term
36 months
Save 22%$2,990one-off$83 per month of new files- The full archive to date on day one
- New daily files for 36 months
- Downloaded files stay licensed after the term
Every option includes the full archive to date. A longer term costs less per month of new files.
Need a written offer before you order? Write to us.
What you get
- 19 daysPoint-in-time archive since 3 Nov 2026 · 40,809,391 rows
- 2,626 daysReconstructed history since 10 Sept 2019 · 30,890,711 rows
- Every dayNew files of the datasets stored live, ready by about 00:10 UTC
- JSON linesCSV in every new day, Parquet for part of the days so far · Gzip, schema and README per dataset
969.8 MB point in time, 1.78 GB with the reconstructed history, stored as gzip files with every format counted. Point in time: stored live as measured, one file set per UTC day. Reconstructed: recorded values compiled afterwards, labelled in every file.
Stored live, every day
As measured, from 3 Nov 2026, final once the day is complete. For backtests without look-ahead.
- Liquidation heatmap (real positions)
- Liquidation cluster observations
- Cluster hits
- Whale position changes
- Whale stop / take-profit orders (end of day)
- Hyperliquid positions (end of day)
- Tracked positioning per coin (end of day)
- Hyperliquid liquidations
- Liquidations per 5 minutes (all exchanges we cover)
- Funding, open interest, marks (5 min)
- Long / short positioning (5 min)
Reconstructed history
Recorded values reconstructed afterwards by Blockmetric Research, not modelled. Labelled in every file. Files reach 17 Nov 2026; newer days are added in monthly runs.
- Funding, open interest and marks (history) · from 10 Sept 2019
- Open interest per exchange (history) · from 20 Jul 2020
- Long / short account ratio per exchange (history) · from 20 Jul 2020
- Long / short of all Hyperliquid positions (hourly) · from 27 Sept 2025
- Hyperliquid liquidations (history) · from 25 May 2025
- Liquidation heatmap, daily snapshot of every position · from 15 Dec 2025
- Open positions per coin, daily snapshot of every position · from 15 Dec 2025
Every dataset in detail
Point-in-time datasets
Stored as measured by our collector from 3 Nov 2026, a new file every UTC day, final once the day is complete; a later correction is versioned in the manifest. Use these for backtests without look-ahead. Days, rows and gzip size count the files a licence delivers today.
- Liquidation heatmap (real positions)
heatmap_realLiquidation levels of the tracked positions per snapshot, every 5 minutes by default.Test with it: does price move towards the price levels with large liquidation notional?19 days · 32,269,431 rows · 161.8 MB as JSON lines - Liquidation cluster observations
cluster_observationsLarge liquidation clusters as we observed them, with their distance to the mark.Test with it: how often and how fast was a cluster of a given size and distance reached?19 days · 2,472 rows · 83 kB as JSON lines - Cluster hits
cluster_hitsWhich observed clusters the Hyperliquid mark reached, checked up to 72 hours later.Test with it: time from observation to the touch, per coin, side and distance.19 days · 751 rows · 22 kB as JSON lines - Whale position changes
position_eventsOpens, closes, flips, adds, reductions and liquidations of tracked wallets.Test with it: what happened to price after large wallets opened, added to or closed a position?19 days · 1,753,753 rows · 89.4 MB as JSON lines - Whale stop / take-profit orders (end of day)
whale_stops_eodOpen stop and take-profit orders of the largest tracked wallets at the end of each day.Test with it: do resting stop and take-profit orders of large wallets mark levels that price reacts to?19 days · 1,038 rows · 38 kB as JSON lines - Hyperliquid positions (end of day)
positions_eodEvery tracked Hyperliquid position, captured right after each UTC day ends.19 days · 1,664,518 rows · 114.9 MB as JSON lines - Tracked positioning per coin (end of day)
positions_aggregateThe same positions summed per coin and side: positions, wallets and notional.19 days · 12,175 rows · 289 kB as JSON lines - Hyperliquid liquidations
liquidations_hlHyperliquid liquidations of every wallet from the fill record, tracked wallets confirmed by their own fills.19 days · 294,132 rows · 4.6 MB as JSON lines - Liquidations per 5 minutes (all exchanges we cover)
liquidations_5mLiquidations summed per venue, coin and side every 5 minutes.19 days · 348,893 rows · 4.8 MB as JSON lines - Funding, open interest, marks (5 min)
market_5mMark, hourly funding and open interest per venue and coin every 5 minutes.19 days · 4,135,708 rows · 93.7 MB as JSON lines - Long / short positioning (5 min)
long_short_5mLong / short ratios and the tracked Hyperliquid long / short notional every 5 minutes.19 days · 326,520 rows · 5.3 MB as JSON lines
Reconstructed history datasets
Recorded values, reconstructed afterwards by Blockmetric Research from market and on-chain data: no value is modelled, and a value that was not stored stays empty. Not point-in-time: a file shows what our database held when it was exported, and every file is labelled as reconstructed. Part of your licence as soon as its files exist.
- Funding, open interest and marks (history)
market_historyFunding, open interest and marks per venue and coin, hourly. Hyperliquid, Lighter, Binance and Bybit.Available as files10 Sept 2019 to 17 Nov 2026, 2,626 days - Open interest per exchange (history)
open_interest_historyOpen interest of Binance and Bybit per coin before our own collection, hourly, as value in USDT.Available as files20 Jul 2020 to 19 May 2023, 1,020 days - Long / short account ratio per exchange (history)
long_short_ratio_historyLong/short ratio of the accounts on Bybit per coin, hourly from July 2020, and the account and top-trader ratios of Binance.Available as files20 Jul 2020 to 17 Nov 2026, 1,024 days - Long / short of all Hyperliquid positions (hourly)
long_short_historyEvery open Hyperliquid position summed per coin and side at each hourly funding round.Available as files27 Sept 2025 to 17 Nov 2026, 417 days - Hyperliquid liquidations (history)
liquidations_hlEvery Hyperliquid liquidation from the Hyperliquid fill record and our collector.Available as files25 May 2025 to 17 Nov 2026, 542 days - Liquidation heatmap, daily snapshot of every position
heatmap_realLiquidation levels of every open Hyperliquid position, one snapshot per UTC day.Available as files15 Dec 2025 to 17 Nov 2026, 338 days - Open positions per coin, daily snapshot of every position
positions_aggregateEvery open Hyperliquid position summed per coin and side, one snapshot per UTC day.Available as files15 Dec 2025 to 17 Nov 2026, 338 days
Not in the archive
- Price candles (perps, Hyperliquid USDC spot pairs and oracle reference prices): through the API, not as archive files. API plans
- Per-wallet positions before our own collection: the daily snapshots of every position are kept only as sums per coin.
Columns, types and the first rows of a free sample day of every dataset: Try the data first. The full schema: Archive files.
Try the data first
The rows below are the first lines of the free sample of 11 Nov 2026 (BTC, ETH, HYPE), read from its files. Every file of it downloads without an account, with a README and checksums. A sample week of 7 consecutive days, for testing a time series, is free from 25 Nov 2026. All free samples
Liquidation heatmap (real positions) heatmap_real139,479 rows in the sample
Liquidation levels of the tracked positions per snapshot, every 5 minutes by default.
| Column | Type | Meaning |
|---|---|---|
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
ts | timestamp | UTC time (ISO 8601). |
mark | number | Mark at the snapshot. |
bucket_size | number | Bucket width (price). |
bucket_px | number | Lower edge of the bucket. |
long_usd | number | Long notional liquidated in the bucket (USD). |
short_usd | number | Short notional liquidated in the bucket (USD). |
First 8 rows of 139,479 in the sample of 11 Nov 2026
coin | ts | mark | bucket_size | bucket_px | long_usd | short_usd |
|---|---|---|---|---|---|---|
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 36500 | 261500 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 37000 | 1003874 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 37500 | 3031988 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 38000 | 3075469 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 38500 | 37757273 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 39000 | 1802896 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 39500 | 1772070 | 0 |
| BTC | 2026-11-11T00:00:00.000Z | 73780 | 500 | 40000 | 2345341 | 0 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/heatmap_real.csv.gz", parse_dates=["ts"], storage_options={"User-Agent": "pandas"})Liquidation cluster observations cluster_observations32 rows in the sample
Large liquidation clusters as we observed them, with their distance to the mark.
| Column | Type | Meaning |
|---|---|---|
id | integer | Observation id (joins cluster_hits). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
layer | string | Always real (tracked Hyperliquid positions). |
observed_at | timestamp | Observation time. |
side | string | long | short. |
bucket_px | number | Lower edge. |
bucket_top | number | Upper edge. |
usd | number | Notional in the cluster (USD). |
positions | integer | Positions in the cluster. |
mark | number | Mark at observation. |
distance_pct | number | Percent from the mark to the near edge. |
rank | integer | Rank by size at observation. |
First 8 rows of 32 in the sample of 11 Nov 2026
id | coin | layer | observed_at | side | bucket_px | bucket_top | usd | positions | mark | distance_pct | rank |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1703 | HYPE | real | 2026-11-11T00:02:34.837Z | long | 67.5 | 68 | 30508384 | 43 | 72.191 | 5.805 | 2 |
| 1708 | BTC | real | 2026-11-11T01:03:42.868Z | long | 65000 | 65500 | 98007365 | 82 | 73526 | 10.916 | 1 |
| 1710 | BTC | real | 2026-11-11T01:45:09.090Z | long | 64500 | 65000 | 82342030 | 63 | 73822.5 | 11.951 | 2 |
| 1718 | BTC | real | 2026-11-11T06:15:12.721Z | long | 71500 | 72000 | 64444186 | 178 | 73837 | 2.488 | 3 |
| 1722 | ETH | real | 2026-11-11T08:00:07.990Z | long | 1930 | 1940 | 31128800 | 19 | 2069.1 | 6.239 | 1 |
| 1723 | BTC | real | 2026-11-11T09:15:18.568Z | long | 70500 | 71000 | 79546597 | 127 | 73654.8 | 3.604 | 2 |
| 1725 | BTC | real | 2026-11-11T10:15:15.164Z | long | 64000 | 64500 | 87811661 | 72 | 73699 | 12.482 | 1 |
| 1727 | ETH | real | 2026-11-11T10:30:03.474Z | long | 1940 | 1950 | 27564323 | 24 | 2061.7 | 5.418 | 1 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/cluster_observations.csv.gz", parse_dates=["observed_at"], storage_options={"User-Agent": "pandas"})Cluster hits cluster_hits1 row in the sample
Which observed clusters the Hyperliquid mark reached, checked up to 72 hours later.
| Column | Type | Meaning |
|---|---|---|
id | integer | Observation id. |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
layer | string | Always real (tracked Hyperliquid positions). |
observed_at | timestamp | Observation time. |
side | string | long | short. |
bucket_px | number | Lower edge. |
usd | number | Notional at observation (USD). |
reached_at | timestamp | First time the sampled mark touched the bucket. |
reach_source | string | hl_mark_30s | history_5m. |
First 1 row of 1 in the sample of 11 Nov 2026
id | coin | layer | observed_at | side | bucket_px | usd | reached_at | reach_source |
|---|---|---|---|---|---|---|---|---|
| 1662 | ETH | real | 2026-11-10T16:45:02.156Z | short | 2090 | 22876741 | 2026-11-11T17:17:09.747Z | hl_mark_30s |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/cluster_hits.csv.gz", parse_dates=["observed_at", "reached_at"], storage_options={"User-Agent": "pandas"})Whale position changes position_eventsLicence only
Opens, closes, flips, adds, reductions and liquidations of tracked wallets.
Not in the free sample: this dataset holds wallet addresses. The files come with a licence; the columns are below.
| Column | Type | Meaning |
|---|---|---|
ts | timestamp | UTC time (ISO 8601). |
address | string | Hyperliquid address, lowercase (on-chain state). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
kind | string | open | close | increase | decrease | flip | liquidated (fill-confirmed). |
side | string | long | short. |
size_delta | number | Change in coins. |
usd_delta | number | Change in notional (USD). |
price | number | Mark at detection. |
entry_px | number | Entry price after the change. |
liq_px | number | Liquidation price after the change. |
leverage | number | Leverage after the change. |
Whale stop / take-profit orders (end of day) whale_stops_eodLicence only
Open stop and take-profit orders of the largest tracked wallets at the end of each day.
Not in the free sample: this dataset holds wallet addresses. The files come with a licence; the columns are below.
| Column | Type | Meaning |
|---|---|---|
address | string | Hyperliquid address, lowercase (on-chain state). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
kind | string | stop | take_profit. |
side | string | buy | sell (direction when triggered). |
price | number | Trigger price. |
size | number | Size in coins. |
usd | number | size × price (USD). |
reduce_only | boolean | Reduce-only. |
position_tpsl | boolean | Attached to the whole position. |
attached | boolean | Child of a not yet filled entry order. |
order_type | string | Order type as reported. |
placed_at | timestamp | Placement time. |
fetched_at | timestamp | Read time. |
Hyperliquid positions (end of day) positions_eodLicence only
Every tracked Hyperliquid position, captured right after each UTC day ends.
Not in the free sample: this dataset holds wallet addresses. The files come with a licence; the columns are below.
| Column | Type | Meaning |
|---|---|---|
address | string | Hyperliquid address, lowercase (on-chain state). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
side | string | long | short. |
size | number | Absolute size in coins. |
entry_px | number | Entry price. |
position_usd | number | Notional at mark (USD). |
leverage | number | Leverage. |
leverage_type | string | cross | isolated. |
liq_px | number | Liquidation price reported by Hyperliquid (empty = none). |
margin_used | number | Margin used (USD). |
unrealized_pnl | number | Unrealised PnL (USD). |
updated_at | timestamp | Poll time of this position. |
Tracked positioning per coin (end of day) positions_aggregate6 rows in the sample
The same positions summed per coin and side: positions, wallets and notional.
| Column | Type | Meaning |
|---|---|---|
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
side | string | long | short. |
positions | integer | Number of positions. |
position_usd | number | Summed notional (USD). |
with_liq_px_usd | number | Notional of positions that have a liquidation price (USD). |
newest_poll | timestamp | Newest poll time of the group. |
First 6 rows of 6 in the sample of 11 Nov 2026
coin | side | positions | position_usd | with_liq_px_usd | newest_poll |
|---|---|---|---|---|---|
| BTC | long | 5136 | 1442044789.735259 | 1337580988.044836 | 2026-11-12T00:52:24.263Z |
| BTC | short | 2418 | 1467918277.133352 | 1467918277.133352 | 2026-11-12T00:52:24.352Z |
| ETH | long | 2656 | 783599939.301294 | 727350421.292974 | 2026-11-12T00:52:23.026Z |
| ETH | short | 1435 | 790808735.270829 | 790808735.270829 | 2026-11-12T00:52:24.110Z |
| HYPE | long | 3189 | 756667286.621951 | 677781974.871937 | 2026-11-12T00:52:24.263Z |
| HYPE | short | 1463 | 767524731.895185 | 767524731.895185 | 2026-11-12T00:52:24.352Z |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/positions_aggregate.csv.gz", parse_dates=["newest_poll"], storage_options={"User-Agent": "pandas"})Hyperliquid liquidations liquidations_hl1,656 rows in the sample
Hyperliquid liquidations of every wallet from the fill record, tracked wallets confirmed by their own fills.
| Column | Type | Meaning |
|---|---|---|
ts | timestamp | UTC time (ISO 8601). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
side | string | long | short. |
price | number | Price. |
qty | number | Size in coins. |
usd | number | Notional (USD). |
First 8 rows of 1,656 in the sample of 11 Nov 2026
ts | coin | side | price | qty | usd |
|---|---|---|---|---|---|
| 2026-11-11T00:57:47.188Z | ETH | long | 2048.4 | 6.847 | 14025.394800000002 |
| 2026-11-11T01:29:30.883Z | ETH | short | 2056.7 | 2.2 | 4524.74 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.1565 | 323.75154999999995 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.0159 | 32.89233 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.6419 | 1327.89853 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.1955 | 404.43084999999996 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.0106 | 21.92822 |
| 2026-11-11T01:41:29.638Z | ETH | short | 2068.7 | 0.1072 | 221.76463999999999 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/liquidations_hl.csv.gz", parse_dates=["ts"], storage_options={"User-Agent": "pandas"})Liquidations per 5 minutes (all exchanges we cover) liquidations_5m891 rows in the sample
Liquidations summed per venue, coin and side every 5 minutes.
| Column | Type | Meaning |
|---|---|---|
bucket | timestamp | Start of the 5-minute bucket (UTC). |
venue | string | hyperliquid | binance | bybit | okx (enabled sources only). |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
side | string | long | short. |
count | integer | Number of liquidations. |
usd | number | Summed notional (USD). |
First 8 rows of 891 in the sample of 11 Nov 2026
bucket | venue | coin | side | count | usd |
|---|---|---|---|---|---|
| 2026-11-11T00:10:00.000Z | binance | BTC | long | 1 | 957.2134 |
| 2026-11-11T00:10:00.000Z | binance | ETH | short | 1 | 102536.99999999999 |
| 2026-11-11T00:20:00.000Z | binance | BTC | long | 3 | 9782.957499999999 |
| 2026-11-11T00:25:00.000Z | binance | BTC | long | 5 | 16757.6195 |
| 2026-11-11T00:25:00.000Z | bybit | BTC | long | 2 | 4982.4256 |
| 2026-11-11T00:30:00.000Z | binance | ETH | short | 1 | 33076.645 |
| 2026-11-11T00:30:00.000Z | bybit | HYPE | long | 2 | 39.320969999999996 |
| 2026-11-11T00:35:00.000Z | binance | BTC | short | 1 | 2212.029 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/liquidations_5m.csv.gz", parse_dates=["bucket"], storage_options={"User-Agent": "pandas"})Funding, open interest, marks (5 min) market_5m1,719 rows in the sample
Mark, hourly funding and open interest per venue and coin every 5 minutes.
| Column | Type | Meaning |
|---|---|---|
venue | string | Venue. |
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
ts | timestamp | UTC time (ISO 8601). |
mark | number | Mark. |
funding_hourly | number | Funding per hour (fraction). |
open_interest_usd | number | Open interest (USD). |
mark_high | number | Highest sampled mark. |
mark_low | number | Lowest sampled mark. |
First 8 rows of 1,719 in the sample of 11 Nov 2026
venue | coin | ts | mark | funding_hourly | open_interest_usd | mark_high | mark_low |
|---|---|---|---|---|---|---|---|
| hyperliquid | BTC | 2026-11-11T00:00:00.000Z | 73744 | 0.0000125 | 2998912259.42176 | 73780 | 73744 |
| hyperliquid | BTC | 2026-11-11T00:05:00.000Z | 73704 | 0.0000125 | 2995774661.9779196 | 73772 | 73704 |
| hyperliquid | BTC | 2026-11-11T00:10:00.000Z | 73653 | 0.0000125 | 2991926005.66308 | 73681 | 73653 |
| hyperliquid | BTC | 2026-11-11T00:15:00.000Z | 73653 | 0.0000125 | 2994242145.039 | 73653 | 73651 |
| hyperliquid | BTC | 2026-11-11T00:25:00.000Z | 73580 | 0.0000125 | 2993953578.2780004 | 73610 | 73580 |
| hyperliquid | BTC | 2026-11-11T00:30:00.000Z | 73669 | 0.0000125 | 3002417524.74552 | 73669 | 73515 |
| hyperliquid | BTC | 2026-11-11T00:35:00.000Z | 73706 | 0.0000125 | 3003016332.7239203 | 73706 | 73683 |
| hyperliquid | BTC | 2026-11-11T00:40:00.000Z | 73588.9 | 0.0000125 | 2998488438.9666357 | 73725 | 73588.9 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/market_5m.csv.gz", parse_dates=["ts"], storage_options={"User-Agent": "pandas"})Long / short positioning (5 min) long_short_5m838 rows in the sample
Long / short ratios and the tracked Hyperliquid long / short notional every 5 minutes.
| Column | Type | Meaning |
|---|---|---|
coin | string | Coin as named on Hyperliquid (BTC, ETH, kPEPE, HIP-3 names like xyz:TSLA). |
ts | timestamp | UTC time (ISO 8601). |
binance_account | number | Binance long/short account ratio. |
binance_top_account | number | Binance top traders, accounts. |
binance_top_position | number | Binance top traders, positions. |
bybit_account | number | Bybit account ratio. |
okx_account | number | OKX account ratio. |
hl_long_usd | number | Tracked Hyperliquid long notional (USD). |
hl_short_usd | number | Tracked Hyperliquid short notional (USD). |
hl_long_accounts | integer | Tracked Hyperliquid wallets long. |
hl_short_accounts | integer | Tracked Hyperliquid wallets short. |
First 8 rows of 838 in the sample of 11 Nov 2026
coin | ts | binance_account | binance_top_account | binance_top_position | bybit_account | okx_account | hl_long_usd | hl_short_usd | hl_long_accounts | hl_short_accounts |
|---|---|---|---|---|---|---|---|---|---|---|
| BTC | 2026-11-11T00:00:00.000Z | - | - | - | - | - | 1458539392 | 1488767377 | 5226 | 2395 |
| BTC | 2026-11-11T00:15:00.000Z | - | - | - | - | - | 1456121376 | 1485734261 | 5237 | 2387 |
| BTC | 2026-11-11T00:35:00.000Z | - | - | - | - | - | 1460384983 | 1490508582 | 5239 | 2391 |
| BTC | 2026-11-11T00:45:00.000Z | - | - | - | - | - | 1460331349 | 1489590878 | 5239 | 2395 |
| BTC | 2026-11-11T00:55:00.000Z | - | - | - | - | - | 1457450251 | 1487389295 | 5247 | 2394 |
| BTC | 2026-11-11T01:10:00.000Z | - | - | - | - | - | - | - | - | - |
| BTC | 2026-11-11T01:15:00.000Z | - | - | - | - | - | 1460850503 | 1490476238 | 5246 | 2396 |
| BTC | 2026-11-11T01:20:00.000Z | - | - | - | - | - | 1462736849 | 1491240416 | 5254 | 2392 |
import pandas as pd
df = pd.read_csv("https://blockmetric.io/archive/sample/2026-11-11/long_short_5m.csv.gz", parse_dates=["ts"], storage_options={"User-Agent": "pandas"})Example notebooks
Each notebook answers one question with the free sample files, online or from a downloaded sample folder. One sample day is an illustration, not a result.
How often did price reach a large liquidation cluster, and how fast?
Joins cluster observations with cluster hits and counts the share reached within 24 hours by coin, side and distance.
Notebook (.ipynb)cluster_observationscluster_hitsHow did the long and short notional of tracked positions move against price?
Aligns the tracked long and short notional with the 5-minute mark and compares their changes per coin.
Notebook (.ipynb)long_short_5mmarket_5mWhere did liquidation levels sit relative to price, and how much was liquidated that day?
Sums heatmap notional by distance from the mark and sets the levels inside the day range against the liquidations.
Notebook (.ipynb)heatmap_realliquidations_hlmarket_5m
Files or API
Files: Complete archive
For teams, backtests and research on the whole history.
- Every stored day as files, back to 10 Sept 2019, including history older than the 3 years a plan reads · 2,626 reconstructed and 19 point-in-time day files
- Point-in-time states of each past day: positions and whale stop orders at the end of the day, which the API serves for the current day
- The whole market as ready daily files in bulk, without the API's request limits, as JSON lines (CSV is being added to the earlier days)
- Internal use in one organisation, any number of its employees; files downloaded during the term stay licensed without a time limit
- $1,290 one-off for 12 months of new files
API: Desk or API Pro
For one person or one service that asks for the newest values and the plan's history, monthly or yearly.
- Desk: history everything stored, up to 3 years through the key · $79 a month
- API Pro: history everything stored, up to 3 years through the key · $349 a month
- The same market, liquidation, heatmap and long/short values per request within the plan's limits, as JSON, CSV or NDJSON
- Live values, order flow and price candles, which are not in the files; showing data in your own product needs an API plan. API plans
How delivery works
- Your accountEvery file under Datasets, grouped by dataset and month, with size and SHA-256. Nothing is sent by e-mail.
- Personal tokenOne download token per licence, shown once, revocable. Scripts send it as a Bearer header.
- ScriptsCopy-ready bash, PowerShell and Python that download everything, resume and skip what you already have.
- ChecksumsSHA256SUMS and a JSON file list per licence, for both formats or for one; every day's manifest is public before you buy.
- FormatsDaily files as JSON lines for code. CSV for Excel and Google Sheets is in every new day and is being added to the earlier days. Compressed with gzip: a double click unpacks a file on a Mac, on Windows the free 7-Zip does it. Python and DuckDB read the compressed file directly. Days that already have a Parquet copy offer it on Archive files next to the other formats.
What the licence allows
- Internal research and reports
- Backtesting
- Model training
- Internal dashboards
- Charts, statistics and papers that do not let anyone rebuild the data
- Storage with your cloud provider
- Passing on the files or values that can be rebuilt from them
- Resale or a data product built from the files
- Showing the data to third parties in a website, app, bot or feed (API Pro or Enterprise)
In brief
- Who may use it
- One company, any number of its employees and the contractors working for it.
- After the term
- Files downloaded during the term stay licensed for internal use without a time limit.
- Renewal
- None. The licence does not renew.
- Invoice
- Issued by the payment provider.
Questions
Dataset licences
What exactly do I buy?
A dataset licence for the complete archive: every dataset we archive, with all the history we have on day one, plus every new day for the term you choose (12 months, 24 months or 36 months). One payment, no subscription and no automatic renewal.
How far back does the data go?
Everything we have, in two kinds: point-in-time files, stored live by our collector and final once the day is complete (a later correction is versioned in the manifest), and reconstructed history, recorded values compiled afterwards (no value is modelled) and labelled as such in every file. The start of each dataset is under What you get, above.
Can I test the data first?
Yes. The free samples are the first day of each month for BTC, ETH and HYPE, offered once the day is 7 days old: the 8 datasets without per-wallet data, cut from the files a licence delivers for that day, as CSV and NDJSON with a manifest and SHA256SUMS. Where the first day is not delivered in full, the first day of the month that is stands in. For evaluation and research; production use needs a licence.Free samples
What may I do with the data?
Use it inside one organisation: research, backtesting, model training and internal dashboards. Charts, statistics and papers are fine as long as nobody can rebuild the data from them. Files you downloaded during the term stay licensed for internal use without a time limit.Dataset licence terms
Can I show the data in my app or website?
Not with a dataset licence. Showing our data to your users needs an API plan: API Dev for up to 1,000 monthly active users, API Pro for up to 50,000; redistribution and white-label need an Enterprise contract.API plansEnterprise
How do I get the files?
In your account under Datasets: every file grouped by dataset and month with its size and SHA-256, a JSON file list, SHA256SUMS and ready-made scripts for bash, PowerShell and Python that use your personal download token. A new day is usually ready by about 00:10 UTC. Nothing is sent by e-mail.
Can I extend later?
Yes. Buy again at any time; the new term starts when your current one ends, so there is no gap and no overlap.
What if a day is missing?
A day the archive does not export during the term extends the term by one day. A day is final once it is complete; any later correction is versioned in the manifest (revisions, checksums). A day the collector missed is never filled in afterwards.
Can I get a refund?
Full refund on request until the first file of the licence is downloaded (the free sample does not count). A day the archive does not export during the term extends the term by one day. If we stop the archive during the term, the unused months are refunded pro rata. Otherwise the payment is not refundable once files were downloaded; a refunded or disputed payment ends access at once.Terms
Payment and invoice?
Blockmetric (operated by Pilecode US LLC) is the seller. Dataset licences are sold to businesses only: the order form asks for your company name. You pay once on the payment page of Stripe, which shows the payment methods available in your country and the final amount. Stripe issues the invoice with your company details. Questions about an order go to support.Contact support
I need redistribution, white-label or delivery to my own storage
That is an Enterprise contract: redistribution, white-label, delivery to your own S3 bucket or SFTP server, an SLA and payment by invoice.Contact us