Choose a Free Stock Quote API
That Fits Your App选择适合应用的
免费股票报价 API

TL;DR核心摘要

What a useful stock quote response includes有用的股票报价响应应包含什么

A quote endpoint usually accepts a ticker symbol and returns a JSON object. The minimum useful response identifies the instrument, price, currency, market session, source timestamp, and whether the value is delayed.

报价端点通常接收股票代码并返回 JSON 对象。最少应包含标的、价格、币种、交易时段、数据源时间戳，以及报价是否延迟。

Quote, trade, last price, and bar are not synonyms报价、成交、最新价和 K 线不是同义词

A quote normally describes displayed buying and selling interest: bid price, bid size, ask price, and ask size. A trade records an executed transaction. “Last price” is often the latest eligible trade selected by the provider, while an OHLCV bar aggregates eligible events over an interval. A snapshot may package fields from several of these records, and those fields may not share one timestamp. Ask the provider exactly which event populates each field before using a response for an alert, spread calculation, chart, or execution decision.

报价通常表示公开显示的买卖意愿，包括买价、买量、卖价和卖量；成交记录实际执行的交易。“最新价”通常是供应商按规则选择的最近一笔有效成交，而 OHLCV K 线是对某个周期内符合条件的事件进行聚合。快照可能把多类记录的字段组合在一起，而且这些字段不一定对应同一时间。把响应用于提醒、价差计算、图表或交易决策前，必须确认每个字段究竟由什么事件产生。

Real-time, delayed, and extended-hours quotes实时、延迟与盘前盘后报价

“Real-time” is not a complete specification. A feed may cover one venue rather than the consolidated market, exclude certain trade conditions, update only selected symbols, or deliver a current last trade without a current bid and ask. A delayed feed can be useful for education and low-frequency research, but its user interface must not imply that the value is live. Store the provider’s delay classification and calculate the age of every event from its source timestamp.

“实时”并不是完整规格。数据流可能只覆盖单一交易场所，而不是全市场综合行情；也可能排除部分成交条件、只更新特定标的，或提供最新成交价却没有同步的买卖报价。延迟数据适合教学和低频研究，但界面不能让用户误以为它是实时行情。应保存供应商的延迟分类，并依据来源时间戳计算每条数据的实际时效。

Compare free stock quote API options比较免费股票报价 API 方案

Alpha Vantage currently says the majority of its free datasets allow up to 25 requests per day and that real-time and 15-minute delayed U.S. market data are premium-only. FMP requires an API key for all documented requests. Massive documents U.S. stock data sourced through exchanges, securities information processors, FINRA facilities, and OTC markets, while actual access depends on plan and entitlement. These facts can change, so record the date you verified a plan instead of copying an old comparison table into production requirements.

Alpha Vantage 目前说明，大多数免费数据集每天最多调用 25 次，而美国市场实时和 15 分钟延迟数据属于付费能力。FMP 文档中的请求均要求 API Key。Massive 的文档说明，其美国股票数据来自交易所、证券信息处理机构、FINRA 设施和 OTC 市场，实际可访问的数据取决于套餐和授权。这些信息可能变化，因此应记录核验套餐的日期，不能把旧对比表直接当成生产规格。

Build checklist: from first request to production从首次请求到生产的接入检查

Make a real FMP quote request发起一个真实的 FMP 报价请求

FMP documents the stable quote endpoint below and requires an API key. The validation is intentionally strict: an HTTP 200 response is not enough if the body is empty, changes shape, or lacks the expected symbol and price. Keep the API key in an environment variable rather than source code.

FMP 文档提供了下面的稳定版报价端点，并要求 API Key。示例会进行严格验证：即使 HTTP 返回 200，如果响应为空、结构变化或缺少预期的股票代码与价格，也不能视为成功。API Key 应保存在环境变量中，而不是写进源代码。

import os
import requests

response = requests.get(
    "https://financialmodelingprep.com/stable/quote",
    params={
        "symbol": "AAPL",
        "apikey": os.environ["FMP_API_KEY"],
    },
    timeout=20,
)
response.raise_for_status()
payload = response.json()

if not isinstance(payload, list) or not payload:
    raise RuntimeError(f"Unexpected quote response: {payload!r}")

quote = payload[0]
if quote.get("symbol") != "AAPL" or quote.get("price") is None:
    raise RuntimeError(f"Missing required fields: {quote!r}")

normalized = {
    "symbol": quote["symbol"],
    "last_price": float(quote["price"]),
    "provider": "Financial Modeling Prep",
    "retrieved_at": response.headers.get("Date"),
}

The provider response may expose additional timestamps or fields depending on the endpoint and plan; preserve the raw payload and map only documented fields. Do not substitute the HTTP Date header for market event time—it is only a fallback retrieval marker in this small example. Production code should use the provider’s source timestamp where available, record timezone and delay, and compare the result with a second source during onboarding.

供应商响应可能根据端点和套餐包含额外时间戳或字段；应保留原始响应，只映射文档明确说明的字段。这个小示例中的 HTTP Date 只能作为获取时间标记，不能替代市场事件时间。生产代码应优先使用供应商提供的来源时间戳，记录时区与延迟，并在接入阶段与第二来源交叉验证。

Design a resilient stock quote service设计可靠的股票报价服务

A production quote service should separate provider access from the application-facing model. The adapter authenticates, calls the endpoint, retains the raw response, and maps fields. The validation layer checks identity, numeric ranges, timestamps, session state, and staleness. A cache protects quotas and absorbs short provider interruptions. The delivery layer returns the normalized quote together with source, delay, market state, and last successful update so the UI can communicate uncertainty instead of showing a plausible but stale number.

生产报价服务应把供应商接入与面向应用的数据模型分离。适配器负责认证、调用端点、保留原始响应和映射字段；验证层检查标的身份、数值范围、时间戳、交易状态与陈旧程度；缓存保护配额并吸收短时供应商中断；交付层返回标准化报价，同时带上来源、延迟、市场状态和最后成功更新时间，让界面能够表达不确定性，而不是展示一个看似合理但已经过期的数字。

REST polling or WebSocket streaming?选择 REST 轮询还是 WebSocket 推送？

REST is usually simpler for occasional symbol lookups, scheduled portfolio checks, server-rendered pages, and low-frequency agents. WebSockets fit continuously updating watchlists or event-driven systems, but require connection recovery, subscription state, sequence handling, heartbeats, backpressure, and entitlement monitoring. Polling a REST quote every second does not automatically reproduce a licensed real-time stream; it can exhaust quotas while still missing intermediate events.

REST 更适合偶尔查询标的、定时检查投资组合、服务端渲染页面和低频 Agent。WebSocket 适合持续刷新的自选列表或事件驱动系统，但需要处理连接恢复、订阅状态、序列号、心跳、消息积压和权限监控。每秒轮询一次 REST 报价并不等于获得了合规的实时数据流，它可能既耗尽配额，又遗漏中间事件。

When to use QVeris for stock quote APIs何时使用 QVeris 查找股票报价 API

• Search for quote, historical-price, fundamentals, and market-status capabilities in one place.
• Inspect capability inputs and outputs before wiring them into an agent.
• Keep provider choice flexible while maintaining auditable tool calls.

• 在一个入口搜索报价、历史价格、基本面和市场状态能力。
• 接入 Agent 前检查能力的输入与输出。
• 保持服务商选择灵活，同时保留可审计的工具调用。

A useful QVeris request should state more than “get a stock price.” Include the security identifier and exchange when known, required fields, acceptable delay, regular or extended session, currency, intended use, and maximum staleness. Inspect the discovered capability before calling it: verify whether the output is a quote, trade, aggregate, or snapshot; which provider supplies it; what authentication it needs; and whether timestamps and provenance are returned. This prevents an agent from selecting a convenient endpoint whose data definition does not match the decision.

有效的 QVeris 请求不能只写“获取股票价格”。应尽量提供证券标识与交易所、必要字段、可接受延迟、常规或延长交易时段、币种、使用目的和最大陈旧时间。调用前先检查发现的能力：确认输出属于报价、成交、聚合 K 线还是快照，数据由哪家供应商提供，需要什么认证，以及是否返回时间戳与来源。这样可以避免 Agent 选择调用方便、但数据定义与决策不匹配的端点。

After the call, keep the tool name, provider, input parameters, raw response, retrieval time, normalized output, and validation result together. If a second call returns a materially different price, compare event time, venue, session, delay, and field definition before declaring one source wrong. QVeris supports capability discovery and auditable routing; it does not provide exchange entitlements, override provider licenses, or guarantee that a quote is suitable for execution, compliance, or customer display.

调用完成后，应把工具名称、供应商、输入参数、原始响应、获取时间、标准化结果和验证结论一起保存。如果第二个来源返回明显不同的价格，应先比较事件时间、交易场所、交易时段、延迟和字段定义，再判断某个来源是否错误。QVeris 可以帮助查找能力，并保留可审计的调用记录，但不会提供交易所数据权限，也不能取代供应商许可；它也不保证某个报价适合交易执行、合规或面向客户展示。

FAQ常见问题

External references外部参考链接