Documentation
Ask on ChatGPT

REST API Introduction

The CryptoHFTData REST API provides direct access to our cryptocurrency market data without requiring the Python SDK. This is ideal for developers using other programming languages or those who prefer direct HTTP requests.

API Endpoint

Base URL: https://api.cryptohftdata.com/v1

Download Endpoint: GET /download

Versioning, Errors & Rate Limits

Use the /v1 base URL for stable integrations. Breaking changes receive a new major path. When an operation is retired, CryptoHFTData publishes migration guidance and sendsDeprecation, Sunset, and successorLink headers at least 180 days before removal.

JSON errors contain a human-readable error and a stable machine-readable code. Anonymous downloads are limited to 60 requests per 60-second window. Read RateLimit,RateLimit-Policy, RateLimit-Limit,RateLimit-Remaining, and RateLimit-Reset; after HTTP 429, wait for the number of seconds inRetry-After before retrying.

Data Structure & File Paths

All data is organized in hourly chunks (using UTC time) and stored as compressed Parquet files. Each file contains data for a specific hour of a specific trading pair and data type.

File Path Pattern

{exchange}/{YYYY-MM-DD}/{HH}/{symbol}_{data_type}.parquet

Examples:

  • binance_spot/2025-08-01/20/BTCUSDT_trades.parquet
  • binance_futures/2025-08-01/20/ETHUSDT_orderbook.parquet
  • bybit/2025-08-01/20/BTCUSDT_liquidations.parquet

Data Types Available

Spot Markets

  • orderbook L2 snapshots
  • trades Individual executions
  • ticker 24h statistics

Futures Markets

  • orderbook L2 snapshots
  • trades Individual executions
  • ticker 24h statistics
  • mark_price Mark price updates
  • open_interest Open interest data
  • liquidations Liquidation events

File Format

All data files are stored as Parquet format compressed with Zstandard (zstd) compression. Parquet provides efficient columnar storage with excellent compression ratios and fast query performance.

Processing Required

Downloaded files need to be decompressed (zstd) and then read as Parquet files. Most programming languages have libraries to handle both operations.

Have questions?

Our support team is available to help you with integration.

Contact Support