The Best Free Market Data API
for Each Use Case不同场景下的
最佳免费市场数据 API
Compare leading free options by coverage, freshness, rate limits, licensing, and the job your application actually needs to do.
按覆盖范围、时效性、速率限制、许可与真实任务比较主流免费方案，
找到最适合你的市场数据 API。
Best free market data API: the short answer最佳免费市场数据 API：先看结论
Alpha Vantage combines many datasets in one catalog, but its standard free usage is only 25 requests per day.
Finnhub is a practical starting point for stock, forex, and crypto experiments with a straightforward API.
Twelve Data is useful when one prototype needs to test several markets under a shared credit model. Verify usage rights and real-time entitlements before production.
Massive offers a free Stocks Basic plan with reference data and historical aggregate endpoints for research.
一个目录覆盖多类数据，但标准免费额度目前为每天 25 次请求，更适合低频验证。
股票、外汇与加密市场接口便于上手，适合快速验证行情类产品想法。
适合在统一积分模型下测试多个市场，但正式使用前需核对用途与实时权限。
免费 Stocks Basic 计划包含参考数据与历史聚合接口，适合美股研究原型。
What “market data” means before you compare APIs比较 API 之前，先明确“市场数据”指什么
A market-data API can expose several very different products under one label: end-of-day candles, delayed quotes, live trades, bid and ask prices, order-book depth, instrument reference data, exchange calendars, and corporate actions. Some catalogs also include fundamentals, news, technical indicators, or economic series. The best API is therefore the one that matches the exact decision your application must support—not the provider with the longest endpoint list.
“市场数据 API”可能包含完全不同的数据产品：日终 K 线、延迟报价、实时成交、买卖盘、订单簿深度、证券基础资料、交易日历和公司行为。有些服务还会附带基本面、新闻、技术指标或宏观序列。因此，最佳方案不是端点数量最多的那一家，而是能够准确支撑产品实际任务的那一家。
Quotes and trades
Confirm whether the response contains last trade, bid, ask, size, venue, trade conditions, and a market timestamp. “Current price” is ambiguous unless the documentation states whether it is a live trade, a delayed quote, or the previous close.
Candles and historical bars
Check interval support, session boundaries, time zone, OHLCV definitions, split and dividend adjustments, historical depth, and missing-bar policy. A backtest also needs delisted instruments and point-in-time constituents to avoid survivorship bias.
Order books and streaming
For WebSocket feeds, inspect snapshot-versus-delta semantics, sequence numbers, heartbeat rules, reconnect behavior, and book depth. A REST quote endpoint is not a substitute for an ordered stream when the product must maintain a live book.
Symbols and reference data
Reliable symbol metadata includes exchange identifiers, instrument status, currency, trading sessions, tick size, corporate actions, and mappings for renamed or delisted securities. Ticker symbols alone are not stable identifiers.
Forex, crypto, futures, and options
Multi-asset coverage requires asset-specific checks. Forex needs quote convention and daily cut-off rules; crypto needs venue and weekend behavior; futures need contract and roll definitions; options need expiration, strike, multiplier, and corporate-action handling.
报价与成交
要确认响应中是否包含最新成交价、买价、卖价、数量、交易场所、成交条件和市场时间戳。“当前价格”本身并不够明确，文档必须说明它究竟是实时成交、延迟报价，还是上一交易日收盘价。
K 线与历史行情
需要检查周期、交易时段边界、时区、OHLCV 定义、拆股与分红复权、历史深度以及缺失 K 线的处理方式。用于回测时，还要覆盖退市证券和历史成分股，否则容易产生幸存者偏差。
订单簿与实时推送
使用 WebSocket 时，应检查全量快照和增量更新的含义、序列号、心跳规则、断线重连方式和订单簿深度。如果产品需要维护实时盘口，普通 REST 报价接口并不能替代有顺序保证的数据流。
证券代码与参考数据
可靠的证券资料应包含交易所标识、证券状态、币种、交易时间、最小报价单位、公司行为，以及更名和退市代码映射。Ticker 只是展示代码，并不是永久不变的唯一标识。
外汇、加密资产、期货与期权
多资产覆盖不能只看一个“支持”标记。外汇要确认报价方向和日切规则；加密资产要确认数据来自哪些场所及周末如何更新；期货需要合约和连续合约换月定义；期权则要处理到期日、行权价、合约乘数和公司行为。
The 2026 shortlist by use case2026 候选清单：按场景选择
The list below is deliberately use-case-led. Free plans and exchange entitlements change, so treat it as a starting point and verify the current official pricing, documentation, and terms before shipping.
下面的清单按使用场景组织。免费方案与交易所权限会变化，因此它适合作为起点；上线前仍需核对官方定价、文档与条款。
Useful when you want prices, fundamentals, indicators, forex, or crypto under one key. The official site says most endpoints are free, with a standard limit of 25 requests per day.
A practical candidate for experimenting with stocks, forex, crypto, company data, and streaming workflows. Check the exact exchange coverage and entitlement for every target market.
A flexible choice for testing time series across several asset classes. Its Basic plan is free, while credit weights, market access, and commercial terms determine real capacity.
Its Stocks Basic Free plan is a strong research starting point for U.S. tickers, daily summaries, aggregates, and reference endpoints; real-time and intraday access may require an upgrade.
适合用一个密钥测试价格、基本面、技术指标、外汇或加密数据。官方说明大多数接口可免费使用，标准额度为每天 25 次请求。
可用于股票、外汇、加密资产、公司数据和实时推送实验；每个目标市场都应单独确认交易所覆盖与权限。
适合在多个资产类别上测试时间序列。Basic 计划免费，但实际容量取决于积分权重、市场权限与商业使用条款。
Stocks Basic Free 适合测试美股代码、每日汇总、历史聚合与参考接口；实时和日内数据可能需要升级。
Free market data APIs compared by practical fit按真实适用性对比免费市场数据 API
This comparison separates a useful evaluation tier from a production entitlement. Limits and exchange rights can change without notice, so the provider’s current pricing page and terms are the final authority. Figures below reflect official information reviewed on July 28, 2026.
下面的对比会区分“足以完成评估的免费套餐”和“可以用于正式产品的数据权限”。请求额度与交易所授权可能调整，最终仍应以供应商当前定价和条款为准。表中信息依据 2026 年 7 月 28 日核对到的官方资料整理。
| API | Best fit适合场景 | Useful free boundary免费套餐限制 | Main caution主要注意事项 |
|---|---|---|---|
| Alpha Vantage | Low-frequency experiments across prices, indicators, fundamentals, forex, and crypto低频测试价格、指标、基本面、外汇和加密数据 | Most datasets up to 25 requests per day多数数据集每天最多 25 次请求 | The official support page states that real-time and 15-minute delayed U.S. stock data are premium-only官方说明美国股票实时和延迟 15 分钟数据属于付费内容 |
| Twelve Data | Multi-asset dashboards and time-series prototypes多资产看板和时间序列原型 | Basic plan publishes 8 API credits per minute and 800 per dayBasic 计划公布每分钟 8 个 API 积分、每天 800 个 | Credit cost scales by endpoint and symbol; market rights vary积分消耗会随接口和标的增加，市场授权也不完全相同 |
| Finnhub | Fast experiments with stocks, forex, crypto, company data, and streaming快速测试股票、外汇、加密资产、公司数据和推送 | Developer access is available, subject to endpoint and market restrictions提供开发者访问，但不同接口和市场存在各自限制 | Verify exchange coverage, delay, and commercial rights for every target market每个目标市场都要单独确认交易所覆盖、延迟和商业使用权 |
| Massive | U.S. ticker reference data and historical aggregate research美股代码资料与历史聚合行情研究 | Stocks Basic offers a free starting point; current stock pricing lists five API calls per minuteStocks Basic 提供免费起点；当前股票定价页列出每分钟 5 次 API 调用 | Real-time, intraday, flat-file, and exchange entitlements differ by plan实时、日内、批量文件和交易所授权会随套餐变化 |
| CoinGecko | Crypto asset discovery, market snapshots, and broad token coverage加密资产发现、市场快照和广泛代币覆盖 | Demo access is suitable for evaluation and low-volume requestsDemo 访问适合评估和低频请求 | Aggregated market prices are not executable exchange quotes or order books聚合市场价格不是交易所可成交报价，也不等同于订单簿 |
| Exchange-native APIs | Venue-specific crypto, brokerage, or trading workflows特定交易所、券商或交易流程 | Many venues expose public market endpoints or account-linked access不少交易场所提供公开行情接口或账户绑定访问 | Coverage is narrow and schemas differ; account access does not automatically grant redistribution rights覆盖范围较窄、数据结构不统一，账户可访问也不代表允许再分发 |
Selection rule: use a broad aggregator for discovery and prototyping; move to an exchange, broker, or licensed specialist when the product depends on executable prices, guaranteed latency, deep history, or redistribution.
选择原则：产品发现和原型阶段可以使用综合型数据源；一旦产品依赖可成交价格、稳定延迟、深度历史或数据再分发，就应转向交易所、券商或具备明确授权的专业供应商。
The criteria that separate a demo API from a dependable one区分演示接口与可靠接口的关键标准
| Criterion标准 | What to verify需要确认 | Common free-tier trade-off免费套餐的常见限制 | Test测试方法 |
|---|---|---|---|
| Coverage覆盖 | Assets, exchanges, regions, fields, and history.资产、交易所、地区、字段和历史深度。 | Narrow universe or short history.标的较少或历史较短。 | Query representative and delisted symbols.查询代表性标的和退市标的。 |
| Freshness时效性 | Timestamps, delays, update schedule, and revisions.时间戳、延迟、更新频率和修订规则。 | Delayed or end-of-day data.延迟或仅日终数据。 | Compare several updates with a primary source.将多次更新结果与一手数据源进行对比。 |
| Limits限额 | Per-minute, daily, concurrent, and endpoint quotas.分钟、每日、并发和单接口配额。 | Low burst capacity or hard daily caps.突发容量低或每日硬上限。 | Exercise 429 responses and retry headers.触发 429 并检查重试响应头。 |
| Rights许可与权利 | Commercial use, storage, attribution, and redistribution.商业使用、存储、署名和再分发。 | Personal or non-commercial use only.仅限个人或非商业用途。 | Read both API terms and upstream data rights.同时阅读接口条款与上游数据权利。 |
Which market data API should you choose?不同产品应该选择哪类市场数据 API？
Portfolio tracker or watchlist
You usually need delayed or end-of-day prices, instrument metadata, exchange calendars, currency, and corporate actions. A multi-asset REST API can be sufficient if the UI labels delay clearly and cached updates fit the quota. Validate renamed and delisted symbols before trusting portfolio history.
Charting and technical analysis
Prioritize consistent OHLCV definitions, adjustment factors, session boundaries, time zones, and enough history at each interval. Recalculate one or two indicators locally to verify that the vendor’s adjusted series and candle boundaries match your assumptions.
Backtesting and quantitative research
Historical depth alone is not enough. Look for delisted instruments, corporate actions, point-in-time index membership, correction history, and bulk retrieval. A free endpoint that only returns today’s active universe can produce an unrealistically strong backtest.
Alerts and live monitoring
Confirm whether the feed is truly real time, what event timestamp it carries, how WebSocket reconnects work, and whether missed messages can be replayed. Monitor stale data explicitly: an HTTP 200 response containing yesterday’s price is still an outage for a live alert.
Trading or execution workflows
Use venue- or broker-authorized data when a decision can create an order. Executable bid and ask, market status, sequence integrity, clock synchronization, and service expectations matter more than broad asset coverage. Keep market-data ingestion separate from order submission and risk controls.
AI assistants and financial agents
Agents need more than a price field. Tool descriptions should state the source, timestamp, delay, currency, adjustment method, and allowed use. Require the agent to return provenance with the answer and prevent an estimated or stale value from silently replacing a filed or live value.
投资组合追踪或自选列表
通常需要延迟或日终价格、证券基础资料、交易日历、币种和公司行为。如果界面清楚标注延迟，而且缓存后的刷新频率不超过额度，多资产 REST API 就可能够用。正式使用前要专门测试更名和退市证券，否则历史持仓容易断裂。
图表和技术分析
重点检查 OHLCV 定义是否一致、复权因子、交易时段边界、时区，以及各时间粒度的历史深度。可以在本地重新计算一两个常用指标，用来确认供应商的复权序列和 K 线切分方式是否符合你的假设。
回测和量化研究
历史够长并不代表适合回测。还需要退市证券、公司行为、历史时点指数成分、数据修订记录和批量获取能力。如果免费接口只返回今天仍在交易的标的，回测结果可能因为幸存者偏差而显得异常优秀。
提醒和实时监控
要确认数据是否真正实时、事件时间戳代表什么、WebSocket 如何重连，以及遗漏消息能否补取。系统还应单独监控数据是否过期：接口返回 200 但价格仍停留在昨天，对实时提醒来说仍然是故障。
交易或执行流程
只要数据可能触发订单，就应使用交易场所或券商正式授权的来源。可成交买卖价、市场状态、消息顺序、时钟同步和服务保障比“支持很多资产”更重要。行情采集、订单提交和风险控制也应保持隔离。
AI 助手和金融 Agent
Agent 需要的不只是一个价格字段。工具说明应明确来源、时间戳、延迟、币种、复权方法和使用范围；回答中要保留来源信息，并防止预测值或过期数据在没有提示的情况下替代正式披露或实时数值。
Test the free tier before building around it按免费套餐开发之前，先完成一轮小规模验证
A provider should pass a representative test set before its response model spreads through your application. Use a few symbols chosen to expose edge cases rather than downloading the whole universe immediately.
在应用大量依赖某家供应商的数据结构之前，应先用一组有代表性的样本完成验证。不要一开始就下载整个市场，而应故意选择少量能够暴露边界问题的标的。
1. Define the contract
List required fields, acceptable delay, time zone, adjustment method, units, null behavior, and maximum stale age. This turns “the API works” into a result that can be tested.
2. Build an edge-case set
Include an active liquid symbol, a recent split, a delisted instrument, a market holiday, a missing interval, and a non-U.S. listing if global coverage matters.
3. Reconcile values
Compare timestamps, units, and several observations with an exchange notice, issuer record, or another licensed source. Investigate differences instead of averaging them away.
4. Force failures
Trigger an invalid symbol, 429 response, reconnect, and partial pagination. Record retry headers and verify that the client does not turn an error into a zero price.
5. Model real usage
Calculate symbols × endpoints × refreshes × users, then add retries and backfills. Multi-symbol calls may consume multiple credits even when they look like one request.
6. Review rights
Check internal use, commercial display, redistribution, caching, derived data, attribution, and model training separately. Save the reviewed terms with a date.
1. 写清数据契约
列出必需字段、可接受延迟、时区、复权方法、单位、空值规则和最大允许过期时间。这样才能把“接口能用”变成可验证的标准。
2. 准备边界样本
样本中应包含活跃标的、近期拆股证券、退市证券、市场休市日、缺失时间段；如果需要全球覆盖，还要加入非美国上市证券。
3. 对账返回值
先对齐时间戳和单位，再把若干观测值与交易所公告、公司记录或另一个已获正式授权的数据源比较。出现差异时要找到原因，而不是简单取平均。
4. 主动触发失败
测试无效代码、429 限流、断线重连和分页中断，记录重试响应头，并确认客户端不会误把错误响应转换成零值价格。
5. 估算真实用量
按照“标的 × 接口 × 刷新次数 × 用户量”计算，再加上重试和历史回补。多标的请求看似只有一次，也可能消耗多份积分。
6. 单独审查许可
内部使用、商业展示、再分发、缓存、衍生数据、署名和模型训练应分别核对，并保存审查日期和当时的条款版本。
A safer integration pattern for free market data免费市场数据的稳健接入方法
Map provider-specific symbols, timestamps, fields, and errors into an internal model instead of leaking one response shape through the product.
Use different lifetimes for quotes, candles, exchange calendars, and symbol metadata. Respect terms that restrict storage or redistribution.
Record remaining quota, latency, errors, stale responses, and cache hit rate. Alert before a daily cap becomes a user-visible outage.
Store source, retrieval time, market timestamp, currency, adjustment method, and transformation version with important observations.
把供应商特有的标的代码、时间戳、字段和错误映射为内部数据模型，避免整个产品依赖某一种外部响应格式。
报价、K 线、交易日历与标的元数据应使用不同缓存周期，并遵守对存储或再分发的限制。
记录剩余配额、延迟、错误、过期响应和缓存命中率，在每日上限影响用户前告警。
为重要数据保留来源、抓取时间、市场时间戳、币种、复权方法和转换版本。
Use QVeris to discover market data capabilities用 QVeris 查找市场数据能力
Provider lists go stale and every API describes itself differently. QVeris helps developers and agents discover, inspect, and call relevant capabilities while keeping the evaluation criteria explicit.
供应商列表会过时，不同接口的描述方式也不一致。QVeris 帮助开发者和 Agent 发现、检查并调用相关能力，同时让选型标准保持清晰。
- Search by the capability you need—quotes, trades, candles, order books, exchange reference data, forex, or crypto—not only by a vendor name.
- Inspect inputs, outputs, authentication, and constraints before wiring a capability into an automated workflow.
- Keep provider adapters and provenance in your application even when discovery and calling are streamlined.
- 按报价、成交、K 线、订单簿、交易所参考数据、外汇或加密行情等能力搜索，而不只按供应商名称搜索。
- 接入自动化工作流前，先检查输入、输出、鉴权方式和限制条件。
- 即使查找和调用能力的流程得到简化，也应在应用中保留适配层和来源信息。
FAQ常见问题
There is no universal winner. Alpha Vantage fits low-volume breadth, Finnhub fast prototypes, Twelve Data multi-asset exploration, and Massive U.S. market research.
Sometimes, for selected markets. Verify the exact venue, delay, endpoint, exchange entitlement, and display or redistribution rights.
Sometimes. Confirm licensing, attribution, redistribution rights, quotas, reliability expectations, and an upgrade or fallback path.
Not always, but a provider-neutral data model makes it much easier to add coverage or recover from limits and outages later.
不存在适合所有场景的唯一答案。Alpha Vantage 适合低频广覆盖，Finnhub 适合快速原型，Twelve Data 适合多资产探索，Massive 适合美股研究。
部分市场可以，但必须确认具体交易场所、延迟、接口、交易所权限，以及展示或再分发权。
有时可以，但必须确认许可、署名、再分发权、配额、可靠性预期，以及升级或回退路径。
不一定，但供应商中立的数据模型能让后续扩展覆盖、应对配额变化或服务中断更容易。