CryptoHFTData Documentation

Make your first query in one minute

Install the SDK and paste your API key. The code then gives you a DataFrame. If you are signed in, click Insert Key on the code block to add your key.

pip install cryptohftdata

1import cryptohftdata as chd
2
3client = chd.CryptoHFTDataClient(api_key="your-api-key-here")
4
5df = client.get_trades(
6    symbol="BTCUSDT",
7    exchange=chd.exchanges.BINANCE_SPOT,
8    start_date="2025-08-01",
9    end_date="2025-08-01",
10)
11print(df.head())

If you do not have a key, create a free account. You can also use the REST API without a key. Without a key, the REST API accepts up to 60 requests per minute.

Select an access method

All three methods give the same files. You can use the methods in any combination.

Data types

Each data type has a schema reference. Each data type also has examples that you can run with the Python SDK and with the REST API.

How we organise the data

File layout

One file per symbol, data type and UTC hour

Format

Parquet, compressed with Zstandard (zstd)

Availability

Each hour is available about 15 minutes after it closes

History

From Jun 28, 2025, with new data every hour

File paths use this pattern: {exchange}/{YYYY-MM-DD}/{HH}/{symbol}_{data_type}.parquet. For the full layout, refer to REST API introduction.

Supported exchanges

For each exchange, the data starts on the date in the table. The data continues to the most recent closed hour. If we no longer collect an exchange, the table shows an end date. For the list of known missing hours, refer to Known Gaps & Corrections. For the coverage of each dataset, refer to the dataset pages.

In development: Coinbase, Bitmart.