Free Earnings Transcripts API
Sources Compared免费财报电话会文本 API
来源与限制对比
Free access may mean a preview, selected companies, or a limited request quota. Compare full text, speaker structure, freshness, and rights.
免费可能只代表预览、部分公司或有限配额。应比较全文、发言人结构、新鲜度和使用权。
TL;DR
API Ninjas offers a free key and a structured transcript response with participants, sections, summaries, guidance, and sentiment fields.
Alpha Vantage documents more than 15 years of transcript history with turn-level sentiment, subject to current plan limits.
Finnhub labels transcript retrieval and live earnings-call audio as premium even though other company data has a free tier.
Choose by time-to-publish, speaker segmentation, full-text rights, historical depth, and whether summaries can be traced to passages.
API Ninjas 提供免费 Key，响应包含参与者、分段、摘要、指引和情绪等结构化字段。
Alpha Vantage 文档说明覆盖 15 年以上文本并提供逐轮情绪，但仍需核验当前套餐限制。
Finnhub 虽有其他公司数据免费层，但文本获取和实时电话会音频被明确标为 Premium。
按发布速度、发言人分段、全文权利、历史深度，以及摘要能否追溯到原文选择。
What “free” and “real-time” mean for this API这个 API 中“免费”与“实时”分别意味着什么
A free earnings transcripts API is not simply a text download. Research systems need to know which fiscal quarter the call belongs to, when it occurred, who spoke, whether a passage came from prepared remarks or Q&A, and whether the transcript is complete.
Free offers differ sharply. Some return full text under a request quota, some expose only metadata or a short preview, and some make transcripts a separate premium dataset. The phrase free transcript may also describe a webpage that can be read by a person but cannot legally or technically be automated.
For AI workflows, structure matters as much as coverage. Speaker names, roles, section boundaries, source URLs, publication timestamps, and passage-level citations make it possible to extract guidance without attributing an analyst's question to management.
免费财报电话会文本 API 不只是下载一段文字。研究系统必须知道电话会对应哪个财季、何时举行、谁在发言、内容属于管理层陈述还是问答，以及文本是否完整。
不同免费方案差异很大：有的在配额内返回全文，有的只开放元数据或短预览，有的把文本单独列为 Premium 数据集。所谓免费文本也可能只是人能阅读的网页，并不允许自动化使用。
对 AI 工作流而言，结构与覆盖同样重要。发言人、角色、章节边界、来源 URL、发布时间和段落级引用，能避免把分析师问题误当成管理层指引。
Provider comparison: access, fields, and limits供应商对比：入口、字段与限制
| Source来源 | Free access免费方式 | Structure结构化能力 | Caveat注意点 |
|---|---|---|---|
| API NinjasAPI Ninjas | Free API key; no card stated免费 Key；官方称无需信用卡 | Transcript, participants, sections, summary, guidance, sentiment文本、参与者、分段、摘要、指引、情绪 | Verify current quota and commercial terms核验当前配额与商业条款 |
| Alpha VantageAlpha Vantage | Free key with plan limits免费 Key，受套餐限制 | Quarter transcript and turn-level sentiment按季度文本与逐轮情绪 | Rate limits and coverage require testing请求限制与覆盖需实测 |
| EarningsCalls.devEarningsCalls.dev | Metadata, preview, and summaries on free tier免费层提供元数据、预览和摘要 | Recent feed and structured components最新文本流与结构组件 | Full transcript is paid全文属于付费层 |
| FinnhubFinnhub | Other datasets have free access其他数据集有免费入口 | Transcript, participants, audio, live call fields文本、参与者、音频与实时电话会字段 | Transcript and live call endpoints are premium文本与实时电话会端点为 Premium |
Provider-by-provider notes: where the free tier stops逐供应商深挖：免费层在哪里结束
API Ninjas is a useful evaluation source because its example response goes beyond raw text: participants, roles, prepared remarks versus Q&A, summaries, guidance, risk factors, and sentiment are represented directly. The next step is to test quota, company coverage, transcript completeness, publication delay, and commercial terms with a representative quarter.
Alpha Vantage documents long transcript history and sentiment at the conversational-turn level. That can accelerate research prototypes, but a sentiment field should never replace the original passage. Benchmark speaker attribution, fiscal-quarter mapping, and response limits before using the score in a ranking model.
EarningsCalls.dev illustrates a common free-tier design: recent-call metadata, summaries, components, and a short preview can be available without full transcript access. It is useful for discovery and queue building, but a 250-character preview cannot support comprehensive call analysis or negative claims about topics management did not discuss.
Finnhub documents transcripts, participants, audio, and live earnings-call media with broad historical coverage, while clearly labeling these endpoints premium. It is a candidate when audio alignment and international coverage matter, but it should not be presented as a free transcript source simply because Finnhub has a separate free plan for other datasets.
API Ninjas 的示例响应不只有原始文本，还直接提供参与者、角色、管理层陈述与问答、摘要、指引、风险和情绪，适合评估。下一步应使用代表性财季实测配额、公司覆盖、完整度、发布延迟和商业条款。
Alpha Vantage 文档说明覆盖较长历史，并提供对话轮次级情绪，可加速研究原型。但情绪字段不能替代原文；在用于排序模型前，要评测发言人归因、财季映射和响应限制。
EarningsCalls.dev 展示了常见免费层模式：可以免费获得最新电话会元数据、摘要、组件和短预览，但全文需付费。它适合发现和任务排队，却不能用 250 字符预览做完整分析，更不能断言管理层没有讨论某主题。
Finnhub 文档提供文本、参与者、音频和实时电话会媒体，并有较深历史覆盖，但这些端点被明确标为 Premium。需要音频对齐和国际覆盖时可考虑，但不能因为 Finnhub 其他数据有免费层，就把它写成免费文本源。
The fields that decide whether the data is usable决定数据是否可用的关键字段
Count companies, geographies, history, and newly added calls—not only a familiar ticker demo.
Measure the interval from call end to transcript availability and distinguish draft, machine, and corrected versions.
Validate names, roles, company versus analyst identity, prepared remarks, and Q&A boundaries.
Readable, searchable, downloadable, summarizable, and redistributable are different permissions.
统计公司、地区、历史和新增电话会，而不是只测试一个熟悉的大盘股。
衡量电话会结束到文本可用的时间，并区分草稿、机器转写和校正版。
校验姓名、角色、公司方与分析师身份、管理层陈述和问答边界。
可阅读、可搜索、可下载、可摘要和可再分发是不同权限。
Which option fits your workload?哪一种选择适合你的工作负载？
Metadata, event calendars, recent feeds, and previews can be enough to detect that a new call is available and enqueue downstream work.
Require full text, reliable speaker roles, prepared-remarks/Q&A separation, fiscal metadata, and passage citations for every extracted target or range.
Version text, normalize speakers and KPIs, and preserve both fiscal and calendar dates so comparisons do not drift across reporting calendars.
Return cited passages, source URL, provider, transcript version, speaker, and section. The final prose should distinguish management statements from analyst questions.
元数据、事件日历、最新文本流和预览足以发现新电话会并将其加入下游任务队列。
需要全文、可靠发言人角色、管理层陈述/问答分离、财季元数据，以及每个目标值或区间的段落引用。
对文本做版本管理，统一发言人与 KPI，同时保留财季和日历日期，避免跨报告日历比较时错位。
返回引用段落、来源 URL、供应商、文本版本、发言人和章节，并在最终表述中区分管理层观点与分析师问题。
A practical evaluation and implementation playbook可执行的评估与落地方法
Include accents, multiple executives, operator transitions, guidance revisions, and long Q&A sessions.
Check speaker and section accuracy first. Sentiment is misleading when the wrong person owns the sentence.
Every extracted guidance, risk, KPI, or quote should point to the original transcript segment and call metadata.
Version transcript text and rerun downstream extraction when a corrected transcript replaces an early version.
覆盖不同口音、多位高管、主持人切换、指引修订和长问答。
先检查发言人和章节准确度；句子归属错误时，情绪分数没有意义。
每个指引、风险、KPI 或引用都应指向原始段落和电话会元数据。
当校正版替换早期版本时，对文本做版本管理并重新运行下游抽取。
Common mistakes that make a free API expensive让免费 API 变昂贵的常见错误
A release contains reported numbers; a call adds management explanation and analyst questions.
A free snippet cannot support claims about what management did not discuss.
Calendar dates and fiscal quarters can differ, especially across non-calendar fiscal years.
API access does not automatically grant broad republication rights.
业绩公告给出数字，电话会还包含管理层解释与分析师提问。
短预览不能支持“管理层没有讨论某事”这类结论。
日历日期和财季可能不同，非自然年度公司尤其如此。
能通过 API 获取，不等于拥有广泛再发布权。
Where QVeris fits in the workflowQVeris 在工作流中的位置
QVeris is a capability discovery and routing layer for agents. It can help an agent find, inspect, and call financial capabilities, but it does not replace the underlying data provider's license, entitlement, or source-of-truth obligations.
QVeris 是面向 Agent 的能力发现与路由层，可帮助 Agent 找到、检查并调用金融能力，但不会替代底层数据供应商的许可、权限或真值来源责任。
- Use QVeris to discover transcript or filings capabilities and inspect the expected schema before the agent calls them.
- Attach ticker, fiscal year, fiscal quarter, event time, provider, version, speaker, and section to every extracted claim.
- Use the original provider's terms for full-text access and publication; QVeris does not create transcript redistribution rights.
- 可用 QVeris 发现电话会文本或公告能力，并在 Agent 调用前检查预期结构。
- 每条提取结论都应附带代码、财年、财季、事件时间、供应商、版本、发言人和章节。
- 全文访问与发布仍以原供应商条款为准；QVeris 不创造文本再分发权。
FAQ
Yes, but free can mean a quota, selected coverage, metadata, summaries, or previews. Confirm whether full text is included.
It varies by source and may range from near-live machine text to later corrected versions. Measure it with calls you actually cover.
Yes if the system preserves speaker attribution, separates prepared remarks from Q&A, cites passages, and labels uncertainty.
Not automatically. Models, segmentation, and speaker attribution differ, so benchmark on a common labeled set.
有，但免费可能是有限配额、部分覆盖、元数据、摘要或预览，必须确认是否包含全文。
不同来源从接近实时的机器稿到稍后校正版不等，应使用目标公司实测。
可以，前提是保留发言人归因、区分管理层陈述与问答、引用段落并标注不确定性。
不能。模型、分段和发言人归因不同，应在统一标注集上评测。
Sources, methodology, and update policy来源、方法与更新政策
This guide compares public documentation and evaluates providers by workload, data freshness, structure, delivery, legal use, and upgrade path. It does not treat a marketing label as proof of real-time coverage.
Provider features and commercial terms can change. Official documentation linked below is the source of truth; run a representative data test and obtain appropriate rights before production deployment.
本文依据公开文档，按工作负载、数据新鲜度、结构、交付、合法用途和升级路径比较供应商，不把营销标签当作实时覆盖的证明。
供应商功能和商业条款可能变化。以下官方文档是最终依据；生产部署前应运行代表性数据测试并取得适当权利。