Early access is open. Free sample datasets are live now. Download a sample

Data product · Early access

Crypto Order Book Data: L2 Depth History

How much size is resting near the price, on each side, at any moment, so you can estimate slippage before you trade and study how liquidity behaves around big moves.

Coverage

What you get

Item Detail
Product code CR-BOOK
Coverage Spot and perpetual books on the crypto exchanges we carry
Record types Top of book, L2 snapshots (25 levels), and depth bands (cumulative size within 1% to 5% of mid). Full incremental L2 on Enterprise
Granularity Exchange timestamps, with our receive time alongside
History Full history for top of book and depth bands on Pro; L2 snapshot history on Business
Delivery WebSocket for live books; Parquet files for history
Update speed Live on Pro and Business

Bars tell you where price went. The book tells you what it would have cost to push it there. Full L2 history is heavy, so we sell it two ways: depth bands for anyone who needs "how much size within 1%" as a time series, and raw snapshots by the gigabyte for researchers who want the levels themselves.

Schema

Schema

Field Type Required Description
symbol string yes Pair, e.g. BTC-USDT
exchange string yes Venue code
timestamp string yes Exchange time, ISO 8601 UTC
received_timestamp string optional Our receive time
record_type string yes snapshot or update
sequence integer optional Exchange sequence for gap checks
bids array yes List of [price, size], best first
asks array yes List of [price, size], best first
depth integer optional Levels included in a snapshot (5, 25, full)

Depth-band records (as in our free sample) carry symbol, timestamp, distance_from_mid_pct, side, cumulative_size and cumulative_notional_usd.

Example snapshot record (illustrative):

{
  "symbol": "BTC-USDT",
  "exchange": "binance",
  "timestamp": "2026-11-01T00:00:01.120Z",
  "received_timestamp": "2026-11-01T00:00:01.124Z",
  "record_type": "snapshot",
  "sequence": 58219773104,
  "bids": [[77581.99, 1.204], [77581.50, 0.350], [77580.00, 2.118]],
  "asks": [[77582.00, 0.912], [77582.40, 0.075], [77583.10, 1.640]],
  "depth": 3
}

Sample data

Sample data

Bitcoin Perpetual Order Book Depth. Cumulative bid and ask depth within 1% to 5% of mid, in coins and USD. First 10 rows:

symbol timestamp distance_from_mid_pct side cumulative_size cumulative_notional_usd
BTC-USDT-PERP 2026-11-01T00:00:01Z -5 bid 10138.132 771074735.2
BTC-USDT-PERP 2026-11-01T00:00:01Z -4 bid 9501.002 723874082.08
BTC-USDT-PERP 2026-11-01T00:00:01Z -3 bid 7865.959 601520050.77
BTC-USDT-PERP 2026-11-01T00:00:01Z -2 bid 6015.362 461632878.88
BTC-USDT-PERP 2026-11-01T00:00:01Z -1 bid 2817.967 217317556.75
BTC-USDT-PERP 2026-11-01T00:00:01Z -0.2 bid 522.381 40450511.3
BTC-USDT-PERP 2026-11-01T00:00:01Z 0.2 ask 589.675 45760508.91
BTC-USDT-PERP 2026-11-01T00:00:01Z 1 ask 2682.932 209014668.23
BTC-USDT-PERP 2026-11-01T00:00:01Z 2 ask 5818.703 455408530.67
BTC-USDT-PERP 2026-11-01T00:00:01Z 3 ask 6573.527 515324604.88

Source: Binance

Reading the first snapshot: within 1% below mid there were about 2,818 BTC of bids, and within 1% above mid about 2,683 BTC of asks.

Use cases

Use cases

Slippage estimates. Before a large market order, look up cumulative_notional_usd at the 1% band on the side you'll hit. If your order is a big share of it, split it or use limits.

Book imbalance signal. Divide bid cumulative_size by ask cumulative_size at the same distance_from_mid_pct. A sustained imbalance shows one side has more resting interest, which some short-term models use as a lean.

Liquidity around events. Track the 0.2% bands through a CPI release or a large liquidation. Depth thinning out just before the move tells you how fragile the book was.

Microstructure research. On Business, rebuild the top 25 levels from bids and asks snapshots, check sequence for gaps, and measure queue sizes and spread changes over time.

Code

Code examples

Example: latest depth bands for the BTC perpetual.

curl "https://api.marketdata.co/v1/crypto/book/depth-bands?symbol=BTC-USDT-PERP&exchange=binance" \
  -H "Authorization: Bearer $MARKETDATA_API_KEY"

Example: Python, bid/ask imbalance at 1% from the free sample.

import pandas as pd

d = pd.read_csv("btc-perp-depth-bands.csv", parse_dates=["timestamp"])
one = d[d["distance_from_mid_pct"].abs() == 1]
wide = one.pivot(index="timestamp", columns="side", values="cumulative_size")
wide["imbalance"] = wide["bid"] / wide["ask"]
print(wide.head())

Pricing

Pricing

Plan Order books
Pro, $79/mo Live top of book and 1% to 5% depth bands, full history
Business, $249/mo L2 snapshots (25 levels), live and history
Enterprise, quote Full incremental L2

Dataset: $1 per GB of compressed Parquet, with a size estimate before checkout; depth bands at $10 per pair per year. Among the providers we studied, free tiers exist, and paid plans run from $30 to $2,000 a month, median about $417.

Side by side

How we compare

8 of the 50 providers we studied sell crypto order books. Under $100 a month you usually get top of book or aggregated depth only, and real L2 history tends to start at several hundred dollars a month. We sell depth bands on Pro, L2 snapshots on Business, and raw history by the gigabyte, so a researcher who needs one pair for one month pays for one pair for one month and keeps the file.

FAQ

Frequently asked questions

What is a depth band?

It is the total resting size between mid and a given distance from it, for example everything on the bid within 1% below mid. That turns a full book into a few numbers you can track over time.

Why do you store two timestamps?

timestamp is when the exchange says the book changed, and received_timestamp is when we got it. The gap shows network delay and helps when you align books with trades.

How do I know if updates are missing?

Check sequence. Exchanges number book updates, so a jump in the sequence means a gap, and you should resync from the next snapshot.

How big are L2 files?

It depends on the pair and how busy the market was. You'll see a GB estimate for your exact selection before you pay.

Can I use the data commercially?

Commercial use follows the plan terms, with commercial internal use from Business. For the free Binance sample, check Binance's terms before commercial display.

Keep reading