Free Futures Data API
Sources, Limits & Examples免费期货数据 API
数据源、限制与接入示例

TL;DR摘要

A useful free futures data API is not defined by price access alone. It must identify the exact contract, venue, trading session, price type, timestamp, and data status. A response for “crude oil” may refer to a specific delivery month, a vendor-created continuous series, a delayed last trade, or an official settlement. Those values solve different problems and should never share one unlabeled field.

真正有用的免费期货数据 API 不能只提供价格，还必须标明具体合约、交易所、交易时段、价格类型、时间戳和数据状态。“原油价格”可能指某个交割月合约、供应商生成的连续合约、延迟最新成交价，或官方结算价。这些数值解决的问题不同，不能放进同一个没有说明的字段。

Understand futures contracts before choosing an API选择 API 前先理解期货合约

Unlike a stock ticker, a futures symbol usually identifies a product root plus an expiration month and year. The same product trades across several delivery months at once, and each month can have a different price, volume, open interest, bid/ask spread, and last trading date. An API that accepts only a root such as ES or CL may be returning a front month or a synthetic continuous series rather than an exchange-listed contract.

与股票代码不同，期货代码通常由品种根代码、到期月份和年份组成。同一品种会同时交易多个交割月，各月份的价格、成交量、持仓量、买卖价差和最后交易日都可能不同。只接受 ES 或 CL 等根代码的 API，返回的可能是近月合约或合成连续序列，而不是交易所挂牌的具体合约。

Contract symbols and expiration month codes合约代码与到期月份代码

CME explains that display codes commonly combine a product code with a single expiration-month letter and a numeric year. The product code and year length can vary by platform, so save the provider's raw symbol and a normalized contract identity separately.

CME 说明，常见显示代码会把品种代码、单字母到期月份和数字年份组合起来。不同平台的品种代码和年份位数可能不同，因此应分别保存供应商原始代码和标准化合约身份。

Store a contract master, not just price rows保存合约主数据，而不只是价格行

A robust dataset needs product root, full provider symbol, exchange, currency, contract multiplier, tick size, expiration month, first notice date when relevant, last trade date, settlement type, and trading calendar. Physical and cash-settled contracts have different delivery risks. Agricultural, energy, metal, equity-index, interest-rate, FX, and cryptocurrency futures also use different listing cycles and session conventions.

可靠的数据集需要保存品种根代码、完整供应商代码、交易所、币种、合约乘数、最小变动价位、到期月份、适用时的首次通知日、最后交易日、结算类型和交易日历。实物交割与现金结算合约具有不同风险；农产品、能源、金属、股指、利率、外汇和加密货币期货也采用不同的挂牌周期与交易时段。

Use a stable internal contract key such as exchange + product + expiration year-month. Keep every provider mapping versioned because symbols can differ across APIs and can change after a vendor migration. Never infer the full year from one digit without a reference date and contract calendar.

内部可使用“交易所 + 品种 + 到期年月”作为稳定合约键，并对不同供应商的代码映射做版本管理，因为 API 之间代码可能不同，供应商迁移后也可能变化。不能脱离参考日期和合约日历，仅凭一位年份数字推断完整年份。

Which futures data types do you actually need?你真正需要哪些期货数据类型

The best option is the one whose data rights and technical limits fit your application. “Real time” can mean direct exchange data, a vendor-normalized stream, or merely frequent updates. For regulated exchange data, fees and entitlements may still apply even when an API account has a free plan.

合适的接口必须同时满足数据授权与技术限制。“实时”可能指交易所直连、供应商标准化推送，也可能只是高频更新。即使 API 账户有免费套餐，受监管交易所的实时数据仍可能需要单独授权或费用。

REST is appropriate for contract metadata, snapshots, historical backfills, and low-frequency monitoring. WebSocket or another streaming protocol is appropriate when event order and low latency matter. A streaming client must authenticate, subscribe, process heartbeats, detect sequence gaps, reconnect, and request a fresh snapshot before applying later updates. “Connected” does not prove that the book is complete.

REST 适合合约元数据、快照、历史补数和低频监控；当事件顺序和低延迟重要时，应使用 WebSocket 等流式协议。流式客户端需要完成鉴权、订阅、心跳处理、序号缺口检测、断线重连，并在继续应用增量更新前重新请求快照。“已连接”并不代表订单簿完整。

Compare free futures market data API options免费期货数据接口选择清单

What “free futures data” can mean“免费期货数据”可能指什么

Separate permanent public data, delayed exchange pages, developer tiers, sandbox symbols, time-limited credits, and downloadable samples. They are not interchangeable. Public regulatory datasets may remain free but contain positions rather than prices. A commercial free tier may expose prices but restrict contracts, history, calls, display, or redistribution. An exchange web page may be suitable for manual reference while prohibiting automated extraction.

应区分长期公共数据、延迟交易所页面、开发者免费套餐、沙盒环境、限时额度和可下载样本，它们不能互相替代。公共监管数据可能长期免费，但提供的是持仓而不是价格；商业免费套餐可能提供价格，却限制合约、历史、调用次数、展示或再分发；交易所网页可能适合人工参考，却不允许自动抓取。

Write the required rights into the shortlist before testing endpoints. A personal notebook, an internal risk report, a customer chart, an automated strategy, and a resold derived signal can fall under different license categories. Technical access is not a license grant.

测试接口前，就应把所需权利写进候选条件。个人 Notebook、内部风险报告、客户图表、自动化策略和转售衍生信号可能对应不同许可类别。技术上能够访问，并不代表已经取得许可。

Continuous futures data, rolls, and back-adjustment连续期货、展期与复权方法

A continuous futures series is usually created by a data provider or researcher; it is not one exchange-listed instrument. The process selects a contract for each date, chooses when to switch to another expiration, and may adjust historical prices to reduce the jump between contracts. Two valid continuous series for the same product can therefore show different levels and returns.

连续期货序列通常由数据供应商或研究者生成，并不是一个真实挂牌交易的单一工具。构建过程需要为每个日期选择合约、决定何时切换到下一到期月，并可能调整历史价格以减少合约切换跳空。因此，同一品种的两个有效连续序列也可能显示不同价格水平和收益率。

Common roll rules常见展期规则

Adjustment methods change the meaning of price复权方法会改变价格含义

Unadjusted concatenation preserves listed contract prices but introduces jumps at rolls. Difference back-adjustment adds or subtracts the roll gap from earlier history, preserving point changes but moving the historical price level. Ratio adjustment scales earlier history, preserving proportional moves but altering the original contract price. A perpetual weighted series blends two contracts during a roll window and may not equal a tradable instrument on any day.

不复权拼接保留真实挂牌合约价格，但在展期处产生跳空；差值后复权把展期价差加减到更早历史，保留点数变化，却改变历史价格水平；比例复权按比例缩放更早历史，保留比例变化，但同样改变原始合约价格；永久加权序列会在展期窗口混合两个合约，某一天的数值可能不等于任何可交易工具。

How to use a futures data API in Python如何用 Python 调用期货数据 API

Working Python example: CFTC futures positioning API可运行的 Python 示例：CFTC 期货持仓 API

The CFTC Public Reporting Environment offers programmatic access to Commitments of Traders datasets. The example below requests a small page from the Disaggregated Futures Only dataset and validates core fields. It intentionally demonstrates free public positioning data—not real-time futures prices. CFTC guidance says limited tokenless use is generally possible, while identified applications receive a separate request pool; production clients should follow the current API documentation.

CFTC 公共报告环境提供交易商持仓数据的编程访问。下面示例从“分类期货持仓”数据集请求一小页记录，并验证核心字段。它展示的是免费公共持仓数据，而不是实时期货价格。CFTC 指南说明，有限的无 Token 访问通常可用，而标识应用可获得独立请求池；生产客户端应遵循当前 API 文档。

import os
from datetime import datetime, timezone

import requests

url = "https://publicreporting.cftc.gov/api/v3/views/72hh-3qpy/query.json"
params = {"pageNumber": 1, "pageSize": 10}
headers = {}

token = os.getenv("CFTC_APP_TOKEN")
if token:
    headers["X-App-Token"] = token

response = requests.get(url, params=params, headers=headers, timeout=20)
response.raise_for_status()
rows = response.json()

required = {
    "market_and_exchange_names",
    "report_date_as_yyyy_mm_dd",
    "open_interest_all",
}
if not isinstance(rows, list):
    raise ValueError("Unexpected CFTC response type")

clean = []
for row in rows:
    missing = required.difference(row)
    if missing:
        raise ValueError(f"Missing CFTC fields: {sorted(missing)}")
    clean.append({
        "source": "CFTC COT Disaggregated Futures Only",
        "market": row["market_and_exchange_names"],
        "report_date": row["report_date_as_yyyy_mm_dd"],
        "open_interest": int(row["open_interest_all"]),
        "retrieved_at": datetime.now(timezone.utc).isoformat(),
    })

print(clean[:2])

For price integration, apply the same discipline to a licensed provider response: preserve the raw contract symbol, venue, exchange timestamp, provider timestamp, price type, delay status, and entitlement. Validate that bars are sorted, unique, timezone-aware, within the contract's listing dates, and consistent with the selected session template before storing them.

接入价格数据时，应对持牌供应商响应采用同样标准：保留原始合约代码、交易所、交易所时间、供应商时间、价格类型、延迟状态和权限。入库前验证K线已排序、无重复、包含时区、位于合约挂牌期间内，并符合选定交易时段模板。

Futures data quality and backtesting checklist期货数据质量与回测检查清单

Futures backtests fail when valid-looking rows are attached to the wrong contract or economic meaning. Validate the data-generating process before optimizing a strategy.

期货回测常常不是因为数值明显错误而失败，而是看似有效的数据行对应了错误合约或错误经济含义。优化策略前，应先验证数据生成过程。

• Trade date and session: overnight activity can belong to the next exchange trade date even when the local calendar date has not changed.
• Price semantics: last, bid, ask, mark, VWAP, preliminary settlement, and final settlement are different values.
• Volume and open-interest timing: open interest may refer to the previous trading day; CME notes that displayed settlement reports can combine current estimated volume with prior-day open interest.
• Contract lifecycle: exclude dates before listing and after final trading, and account for first notice or delivery risk when selecting a contract.
• Roll bias: define the roll using information available at the time. Do not use final full-day volume to make a decision assumed to occur earlier that day.
• Corrections: preserve raw versions and update status so corrected trades, bars, settlements, or open interest do not silently rewrite results.
• Costs: include spread, commissions, slippage, exchange fees, contract multiplier, roll trades, and any currency conversion.
• Negative and exceptional prices: do not reject observations solely because they violate a hard-coded positive-price assumption.

• 交易日与时段：夜盘活动可能属于下一个交易所交易日，即使本地自然日尚未改变。
• 价格含义：最新价、买价、卖价、标记价、VWAP、初步结算和最终结算都是不同数值。
• 成交量与持仓量时间：持仓量可能对应前一交易日；CME 说明，结算报告可能同时包含当日估算成交量和前一日持仓量。
• 合约生命周期：排除挂牌前和最后交易后的日期，选择合约时还要考虑首次通知日或交割风险。
• 展期偏差：必须使用当时可获得的信息决定展期，不能用全天最终成交量模拟当天更早时点的决策。
• 数据更正：保存原始版本和更新状态，避免更正后的成交、K 线、结算价或持仓量在没有记录的情况下改写结果。
• 成本：纳入买卖价差、佣金、滑点、交易所费用、合约乘数、展期交易和币种转换。
• 负价格与异常行情：不能仅因观测不符合“价格必须为正”的硬编码假设而删除。

Build a reproducible validation sample建立可复现验证样本

Test one liquid front month, one thin deferred month, one expired contract, one roll date, one holiday-shortened session, and one corrected settlement. Compare the API output with the official exchange reference or licensed source allowed by its terms. Save the request, response, timestamp, and expected result so schema or methodology changes become visible in automated tests.

测试一个高流动性近月、一个成交稀疏的远月、一个到期合约、一个展期日、一个节假日缩短时段和一个被更正的结算价。在许可允许的范围内，把 API 输出与官方交易所参考或持牌数据源对照，并保存请求、响应、时间戳和预期结果，让字段或方法变化能够在自动化测试中显现。

Use QVeris to find and inspect market data capabilities用 QVeris 查找并检查行情数据能力

• Search the QVeris Tool Finder for market-data capabilities instead of hard-coding an unverified endpoint.
• Inspect required symbols, parameters, authentication, and output schema before allowing an agent to call a provider.
• Keep licensing, latency, and suitability checks in your application’s approval process.

• 通过 QVeris 工具查找器查找行情数据能力，不要让智能体直接调用未经核验的端点。
• 在允许智能体调用前，检查合约代码、参数、鉴权方式和输出数据结构。
• 数据授权、延迟和适用性仍应保留在应用的审批与风控流程中。

FAQ

External references外部参考链接