Free Index Data API
Sources, Limits & Python免费指数数据 API
行情、历史数据与 Python

Define the exact index series before choosing an API选择指数数据 API 前先确定具体指数序列

The request must identify the exact index family, version, currency, and return convention. A price index, gross total-return index, net total-return index, hedged version, and real-time calculated variant can share a familiar name while producing different levels and histories. Store the publisher's stable identifier and methodology version, not only a display symbol.

请求必须明确指数家族、具体版本、币种和回报口径。价格指数、税前总回报、税后总回报、汇率对冲版本与实时计算版本可能共用熟悉名称，却产生不同点位与历史。应保存发布方稳定标识和方法版本，不能只依赖展示代码。

A practical integration usually starts with a documented S&P 500 or global index identifier, then retrieves a latest level or historical OHLC series in consistent JSON or CSV for a dashboard, research notebook, or agent. The endpoint must expose enough reference metadata to prove that the requested and returned series are the same index variant.

实际接入通常从文档规定的标普 500 或全球指数标识开始，再获取最新点位或历史开高低收序列，以统一 JSON 或 CSV 接入研究笔记、行情看板或 Agent。接口必须提供足够的基础资料，证明请求和返回的是同一个指数版本。

Coverage a stock index data API should provide股票指数数据 API 应覆盖哪些内容

Free market index API sources to evaluate可评估的免费股票指数 API 来源

Use a fixed benchmark set rather than a single famous symbol: one domestic equity price index, its total-return sibling, an international index, a currency-converted or hedged version, and one rebalance date. Compare identifiers, level and OHLC definitions, calendars, return treatment, delay, status, revision, constituent availability, and license.

评估时不要只测试一个知名代码，可固定一组基准：本土股票价格指数、对应总回报版本、国际指数、币种转换或对冲版本，以及一个再平衡日，比较标识、点位与 OHLC 定义、日历、回报处理、延迟、状态、修订、成分可用性和许可。

Use the table as a verification checklist rather than an endorsement. Endpoint access, symbol coverage, delay, and licensing can change, so confirm every required series against the provider's current documentation and a known observation date.

下表是验证清单，并不构成对供应商的背书。端点权限、指数代码覆盖、延迟和许可都可能变化，应使用当前官方文档和已知观测日期逐项核对所需序列。

Validate index returns, rebalances, and corporate actions验证指数回报、再平衡与公司行动

Worked return example: price and total-return indices answer different questions回报示例：价格指数与总回报指数回答不同问题

Assume a price index starts at 1,000 and finishes the year at 1,080. Its price return is 8%. If constituent distributions add 25 index points when reinvested under the methodology, a comparable gross total-return index would end near 1,105, or 10.5% above the start. A net total-return version may finish lower after its stated withholding-tax assumptions.

假设某价格指数年初为 1,000 点，年末为 1,080 点，则价格回报为 8%。如果按照指数方法把成分股分派再投资后增加了约 25 点，对应税前总回报指数可能达到约 1,105 点，即较年初上涨 10.5%。税后总回报版本还会依据既定预提税假设得到更低结果。

The levels are illustrative, but the rule is operational: never splice a price series to a total-return series, and never infer dividend income by subtracting two indices unless publisher, base date, currency, calendar, tax convention, and methodology are compatible. Store the exact series identifier with every observation.

以上点位仅用于说明，但处理规则具有普遍性：不能把价格指数和总回报指数拼接成一条历史，也不能在发布方、基准日、币种、日历、税务约定和方法不兼容时，通过两个指数相减推算分派收益。每条观测都应保存准确的序列标识。

Worked rebalance timeline: announcement is not effective membership再平衡时间线：公告不等于已经成为指数成分

Suppose a provider announces an addition on June 5, determines reference weights using June 14 data, publishes final weights on June 19, and makes the change effective after the June 21 close. A point-in-time constituent query for June 10 should keep the old membership while separately exposing the pending announcement. A portfolio tracking the index normally applies the new membership at the methodology's stated effective moment—not retroactively on June 5.

假设指数发布方在 6 月 5 日宣布调入某证券，使用 6 月 14 日数据确定参考权重，6 月 19 日发布最终权重，并于 6 月 21 日收盘后正式生效。那么，查询 6 月 10 日的历史成分时仍应返回旧名单，同时可以单独展示待生效公告。跟踪指数的组合通常应在方法规定的生效时刻应用新成分，而不能追溯到 6 月 5 日。

Keep announcement, reference, publication, implementation, and first-trading dates as separate fields. Historical research should use the information available by the simulated decision time and include deletions, suspended securities, and identifier changes. Using today's constituent list for earlier dates creates both look-ahead and survivorship bias.

公告日、参考日、发布日期、实施日和首次按新名单交易的日期必须分别保存。历史研究只能使用模拟决策时点已经公开的信息，并保留被删除证券、停牌证券和标识变更。用今天的成分名单回填过去，会同时产生前视偏差和存续偏差。

Corporate actions can preserve economics while changing raw fields公司行动可能保持经济价值，却改变原始字段

Index data API Python example and validation指数数据 API Python 示例与验证步骤

Normalize the provider response first, then validate the exact series identity together with its values. This example rejects a response that silently changes publisher, index variant, currency, or return treatment even when the display symbol looks unchanged.

先把供应商响应归一化，再把具体指数序列身份与数值一同验证。即使展示代码没有变化，只要发布方、指数版本、币种或回报口径发生无提示切换，下面的示例都会拒绝该响应。

from datetime import datetime
from decimal import Decimal, InvalidOperation


def parse_time(value: str) -> datetime:
    parsed = datetime.fromisoformat(value.replace("Z", "+00:00"))
    if parsed.tzinfo is None:
        raise ValueError("index timestamp must include a timezone")
    return parsed


def validate_index_history(payload: dict, expected: dict) -> list[dict]:
    identity = ("publisher", "index_id", "variant", "currency", "return_type")
    mismatched = [key for key in identity if payload.get(key) != expected.get(key)]
    if mismatched:
        raise ValueError(f"unexpected index series: {mismatched}")
    if payload.get("status") not in {"official", "corrected"}:
        raise ValueError("history is not an official or corrected series")

    rows = []
    previous_time = None
    for raw in payload.get("values", []):
        timestamp = parse_time(raw["timestamp"])
        try:
            opening = Decimal(str(raw["open"]))
            high = Decimal(str(raw["high"]))
            low = Decimal(str(raw["low"]))
            close = Decimal(str(raw["close"]))
        except (KeyError, InvalidOperation) as exc:
            raise ValueError("invalid index OHLC row") from exc
        if min(opening, high, low, close) <= 0:
            raise ValueError("index levels must be positive")
        if high < max(opening, close) or low > min(opening, close) or high < low:
            raise ValueError(f"impossible OHLC at {timestamp.isoformat()}")
        if previous_time is not None and timestamp <= previous_time:
            raise ValueError("index history must be strictly increasing")
        rows.append({"timestamp": timestamp, "open": opening,
                     "high": high, "low": low, "close": close})
        previous_time = timestamp
    if not rows:
        raise ValueError("response contains no index observations")
    return rows

How QVeris helps connect index data capabilitiesQVeris 如何连接指数数据能力

• Search separately for index levels, historical series, constituents, weights, and methodology or reference data.
• Require publisher, exact index variant, currency, return treatment, interval, date range, freshness, status, and entitlement in the capability request.
• Preserve provider IDs, source timestamps, revision state, methodology version, and license with every downstream result.

• 分别搜索指数点位、历史序列、成分、权重，以及方法或基础资料能力。
• 在能力请求中明确发布方、具体指数版本、币种、回报处理、周期、日期范围、时效、状态和权限。
• 为每条下游结果保留供应商 ID、来源时间、修订状态、方法版本和许可。

• Open the QVeris provider details for relevant financial-data capabilities.
• Inspect inputs, outputs, authentication, provider documentation, and supported index identifiers before connecting.
• Keep provider, effective date, retrieval time, and license notes with each result; QVeris connectivity is not investment advice.

• 打开 QVeris 服务商详情，查看相关金融数据能力。
• 连接前检查输入、输出、鉴权、供应商文档和支持的指数代码。
• 为每条结果保留供应商、生效日期、抓取时间和许可说明；QVeris 的连接能力不构成投资建议。

Free index data API FAQ免费指数数据 API 常见问题

Authoritative references权威与原始参考链接