预测市场 API11 active endpointsapiKey
Metaculus 数据抓取 API
以规范化 JSON 采集 Metaculus 的问题、社区预测及预测历史、问题元数据、期权和评论动态，用于研究与跟踪。不构成投资、预测或决策建议。
- 公开的 Metaculus 问题、问题详情、预测值、预测历史、元数据和期权
- 在支持范围内提供分类、锦标赛、项目、评论动态及高赞评论端点
- 面向预测研究、公开信号监控和 AI 智能体工作流的结构化 JSON
Structured output
以结构化 JSON 提取 Metaculus 预测数据
使用 Crawlora 的 Metaculus API，无需自建采集器或响应规范化流程，即可获取公开的预测问题、社区预测、预测历史、元数据、期权、分类和锦标赛问题列表以及评论动态。
问题 ID
问题标题
问题 URL
主题或分类
截止日期与解析日期
社区预测
预测历史数据点
预测方法
期权
元数据
锦标赛或项目 slug
评论动态条目
分页元数据
请求元数据
预测研究工作流
用于预测问题监控的 Metaculus API
在作为公开研究数据使用时，公开的 Metaculus 问题和预测可以支撑研究仪表盘、主题监控、锦标赛跟踪、预测历史快照以及 AI 智能体的背景信息。
规划预测市场数据工作流监控社区预测变化
收集预测历史快照
构建分类和锦标赛数据集
查看公开评论和讨论动态
为 AI 研究智能体提供公开预测背景信息
Request schema
Metaculus API 参数
These parameters come from the active Metaculus questions catalog entry.
| Parameter | Type | Required | Description | Example |
|---|---|---|---|---|
| topic | string | No | Optional Metaculus topic slug | - |
| limit | integer | No | Rows to return, default 10, max 25 | - |
Example JSON
Metaculus API 响应示例
This example is rendered from the active endpoint catalog so the page stays aligned with Docs and Playground.
{
"code": 200,
"msg": "OK",
"data": {
"limit": 1,
"source_url": "https://www.metaculus.com/questions/",
"fetched_at": "2026-06-07T14:00:00Z",
"questions": [
{
"id": 43612,
"question_id": 43609,
"title": "Tampa sulphur price in June 2026?",
"status": "open",
"question_type": "numeric",
"forecaster_count": 54,
"comment_count": 16,
"public_page_derived": true
}
]
}
}Endpoint catalog
Metaculus 端点覆盖范围
GETapiKey
/metaculus/questionsMetaculus questions
Returns Metaculus question rows. The endpoint fails closed, returning an error instead of partial data, when the source responds with an authenticated-only response or a challenge page.
Response notes
- `data.questions` contains public page-derived rows, not the authenticated Metaculus API payload. - Forecast values are extracted from the public aggregate payload when present. - Upstream authentication, challenge, or parser drift is returned as a typed upstream error.
MCP tool metaculus_questions
GETapiKey
/metaculus/question/{id}Metaculus question detail
Returns one Metaculus question by numeric `id` (a question or post id). The response has the title, status, question type, unit, open, close and resolve times, forecaster and forecast counts, comment count, latest forecast center and bounds, categories, projects and URL.
Response notes
- The response is derived from public page data, not from the authenticated Metaculus API. - Upstream authentication, challenge, or parser drift is returned as a typed upstream error.
MCP tool metaculus_question
GETapiKey
/metaculus/question/{id}/forecastsMetaculus question forecasts
Returns the latest forecast summary per aggregation method for one Metaculus question, by numeric `id`. Each method has its name, forecaster count, center, lower and upper values, and number of history points. `recency_weighted`, `unweighted` and `single_aggregation` are listed first when present. The question details come with it.
Response notes
- `data.public_page_derived` is `true`. - Preferred methods are ordered first when present: `recency_weighted`, `unweighted`, `single_aggregation`. - Each method exposes forecaster count, the first center/lower/upper value, raw center/bound arrays, and available history point count. - Invalid ids return invalid-parameter errors before upstream IO. - Upstream authentication, challenge pages, missing aggregation data, or parser drift is returned as a typed upstream error.
MCP tool metaculus_question_forecasts
GETapiKey
/metaculus/question/{id}/forecast-historyMetaculus question forecast history
Returns public aggregation forecast history points for one Metaculus question. The `method` enum accepts `recency_weighted`, `unweighted`, and `single_aggregation`.
Response notes
- `data.public_page_derived` is `true`. - Each point exposes timestamp bounds, forecaster count, the first center/lower/upper value, and the raw center/bound arrays. - Invalid ids, methods, or `max_points` values return invalid-parameter errors before upstream IO. - Upstream authentication, challenge, missing history, or parser drift is returned as a typed upstream error.
MCP tool metaculus_question_forecast_history
GETapiKey
/metaculus/question/{id}/metadataMetaculus question metadata
Returns public metadata for one Metaculus question, including option labels, option history, scaling metadata, resolution fields, and timing fields when present.
Response notes
- `data.public_page_derived` is `true`. - `data.options`, `data.all_options_ever`, and `data.options_history` are present for public multiple-choice questions when the page exposes them. - `data.scaling`, `data.resolution`, and `data.possibilities` are returned only when public page data exposes those fields. - Invalid ids return invalid-parameter errors before upstream IO. - Upstream authentication, challenge pages, missing question payloads, or parser drift is returned as a typed upstream error.
MCP tool metaculus_question_metadata
GETapiKey
/metaculus/question/{id}/optionsMetaculus question options
Returns public multiple-choice option labels and latest option-level forecast values for one Metaculus question. The `method` enum accepts `recency_weighted`, `unweighted`, and `single_aggregation`.
Response notes
- `data.public_page_derived` is `true`. - `data.options` aligns public option labels with latest aggregation arrays by index. - Each option may include center, lower, upper, mean, forecast value, and forecaster count when the public payload exposes those fields. - Invalid ids or methods return invalid-parameter errors before upstream IO. - Non-option questions, missing option labels, missing aggregation data, upstream authentication, challenge pages, or parser drift are returned as typed upstream errors.
MCP tool metaculus_question_options
GETapiKey
/metaculus/category/{slug}/questionsMetaculus category questions
Returns Metaculus question rows for a category. Allowed category slugs: artificial-intelligence, computing-and-math, cryptocurrencies, economy-business, elections, environment-climate, geopolitics, health-pandemics, law, metaculus, natural-sciences, nuclear, politics, social-sciences, space, sports-entertainment, technology.
Response notes
- `data.feed` is `category`. - `data.source_url` uses `/questions/?categories={slug}&for_main_feed=false`. - Unsupported category slugs return invalid-parameter errors before upstream IO. - Upstream authentication, challenge, or parser drift is returned as a typed upstream error.
MCP tool metaculus_category_questions
GETapiKey
/metaculus/tournament/{slug}/questionsMetaculus tournament questions
Returns Metaculus question rows for one public tournament, filtered by its slug. A slug that does not exist returns 404.
Response notes
- `data.feed` is `tournament`. - Every returned question belongs to the requested tournament; `data.questions[].projects[].slug` carries it. - A tournament slug that does not exist returns `404`, not an unfiltered feed. - A tournament that exists but currently has no questions returns `200` with an empty `data.questions`. - Invalid slugs or limits return invalid-parameter errors before upstream IO. - Upstream authentication, challenge pages, or parser drift is returned as a typed upstream error.
MCP tool metaculus_tournament_questions
GETapiKey
/metaculus/project/{slug}/questionsMetaculus project questions
Returns Metaculus question rows for one public project, filtered by its slug. A slug that does not exist returns 404.
Response notes
- `data.feed` is `project`. - Every returned question belongs to the requested project; `data.questions[].projects[].slug` carries it. - A project slug that does not exist returns `404`, not an unfiltered feed. - A project that exists but currently has no questions returns `200` with an empty `data.questions`. - Invalid slugs or limits return invalid-parameter errors before upstream IO. - Upstream authentication, challenge pages, or parser drift is returned as a typed upstream error.
MCP tool metaculus_project_questions
GETapiKey
/metaculus/comments-feedMetaculus comments feed
Returns Metaculus question rows for the questions referenced by the most recent public comments, in comment recency order. Comment bodies are not exposed.
Response notes
- `data.feed` is `comments_feed`. - `data.questions` contains the questions referenced by the most recent public comments, ordered by most recent comment first. - Upstream authentication, challenge, or parser drift is returned as a typed upstream error.
MCP tool metaculus_comments_feed
GETapiKey
/metaculus/top-commentsMetaculus top comments feed
Returns Metaculus question rows for the questions whose recent public comments collected the highest vote scores over roughly the last week. Comment bodies are not exposed.
Response notes
- `data.feed` is `weekly_top_comments`. - `data.questions` contains the questions referenced by the last week's public comments, ranked by summed comment vote score. - Upstream authentication, challenge, or parser drift is returned as a typed upstream error.
MCP tool metaculus_top_comments
Managed execution
Crawlora 负责处理 Metaculus 的采集基础设施
Crawlora 将支持的公开 Metaculus 页面和信息流封装为受 API 密钥保护的端点，提供文档化的请求参数、规范化响应、Playground 示例以及基于额度的用量计费。
针对问题、预测、历史、元数据、期权、分类、锦标赛、项目和评论的端点专属请求处理
对支持的公开 Metaculus 页面提供规范化 JSON 响应
Docs 和 Playground 示例均由当前活跃的端点目录生成
以文档化的错误行为取代静默返回不可用的上游响应
基于文档化端点成本的额度计费
一个 Crawlora API 密钥即可调用 Metaculus 及相邻的预测市场端点
Build or buy
Metaculus 抓取方案与自建采集器对比
Use this comparison to decide whether to maintain scraping infrastructure internally or call a managed endpoint.
| Requirement | Building internally | Crawlora |
|---|---|---|
| 端点覆盖范围 | 需要为问题、预测、历史、元数据、分类、锦标赛和评论分别构建采集器。 | 通过一个 API 界面调用文档化的 Metaculus 端点。 |
| 架构规范化 | 需要自行规范化预测、期权、元数据和评论记录。 | 使用当前活跃目录生成的结构化 JSON 字段。 |
| 解析器维护 | 需要跟踪上游页面或信息流的变化，以及空结果的边缘情况。 | 使用 API 背后维护的端点专属采集逻辑。 |
| 开发者测试 | 需要自建示例、文档和测试控制台。 | 接入前先在 Crawlora Docs 和 Playground 中测试。 |
| 用量计费 | 需要自建计量系统和套餐限制。 | 使用基于额度的端点权重和 API 密钥用量跟踪。 |
公开 Metaculus 数据提取，非 Metaculus 官方文档
Crawlora 并非 Metaculus，也不提供 Metaculus 官方 API 文档。Crawlora 为支持的公开 Metaculus 页面提供结构化网络公开数据提取端点。Metaculus 数据仅用于研究和监控工作流，不构成投资、预测、财务、法律或决策建议。客户需自行确保其使用符合适用法律、第三方权利、平台条款及 Crawlora 条款。
Related APIs
相关的预测和市场 API
Connect this endpoint with adjacent Crawlora search, monitoring, docs, and pricing pages.
How to scrape Metaculus
Extract Metaculus data in 4 steps
Crawlora 的 Metaculus 端点是针对无需凭证的公开 Metaculus 页面发起的 GET 请求，遇到已认证的响应或 Cloudflare 验证页面时会直接失败，而不会返回部分数据。列表类：/metaculus/questions 支持 limit 和 topic 参数；/metaculus/category/{slug}/questions 用于获取某个分类；/metaculus/project/{slug}/questions 和 /metaculus/tournament/{slug}/questions 用于获取某个项目或锦标赛；/metaculus/comments-feed 和 /metaculus/top-comments 则返回近期及高赞公开评论所对应的问题。按问题维度：/metaculus/question/{id}、用于获取最新社区综合预测的 /metaculus/question/{id}/forecasts、支持 method（recency_weighted、unweighted 或 single_aggregation）和 max_points 参数的 /metaculus/question/{id}/forecast-history、面向多选题的 /metaculus/question/{id}/options，以及提供解析、量表和时间字段的 /metaculus/question/{id}/metadata。不构成投资、预测或决策建议。
- Step 1
列出你要跟踪的问题
对公开信息流调用 GET /metaculus/questions?limit=<n>&topic=<topic>，或使用 /metaculus/category/{slug}/questions（slug 如 artificial-intelligence、economy-business、elections）、/metaculus/project/{slug}/questions 或 /metaculus/tournament/{slug}/questions 获取指定集合。未知的 slug 会返回 404。
- Step 2
读取当前预测
GET /metaculus/question/{id} 获取规范化的问题数据，GET /metaculus/question/{id}/forecasts 获取按方法划分的最新社区综合预测。对于多选题，/metaculus/question/{id}/options 会返回各选项标签及其最新数值。
- Step 3
获取历史数据
GET /metaculus/question/{id}/forecast-history?method=recency_weighted&max_points=<n> 返回公开综合预测的历史数据点，这正是预测跟踪工具用于绘图的序列。/metaculus/question/{id}/metadata 补充解析、量表和时间字段，便于解读这些数据。
- Step 4
关注讨论动态
GET /metaculus/comments-feed 和 /metaculus/top-comments 返回过去一周内最新及最高赞公开评论所对应的问题，这是判断关注点转移方向的低成本信号。评论正文本身不会被返回。
FAQ
Metaculus API 常见问题
面向正在评估 Crawlora 是否适用于支持的公开 Metaculus 预测页面的开发者的解答。
Crawlora 是否提供 Metaculus API？
是的。Crawlora 为支持的公开 Metaculus 页面提供 Metaculus 端点，并通过文档化的 API 路由返回结构化 JSON。
这是 Metaculus 官方 API 文档吗？
不是。Crawlora 并非 Metaculus，也不提供 Metaculus 官方 API 文档。如需官方平台指南，请参阅 Metaculus 自身的文档；如需 Crawlora 端点参数和响应架构，请查阅 Crawlora Docs。
我可以提取哪些 Metaculus 数据？
当前目录包括公开问题、问题详情、预测值、预测历史、元数据、期权、分类问题列表、锦标赛问题列表、项目问题列表、评论动态记录，以及在支持范围内的高赞评论。
我可以跟踪 Metaculus 社区预测随时间的变化吗？
可以。在支持范围内使用问题预测和预测历史端点，然后按问题 ID 和时间戳存储快照，用于研究仪表盘或监控工作流。
我可以将 Metaculus 数据用于投资或决策建议吗？
不可以。Crawlora 仅返回公开研究数据，不构成投资、预测、财务、法律或决策建议，下游使用责任由客户自行承担。
我需要 Metaculus 的 API 密钥吗？
调用 Crawlora 的 Metaculus 端点时，你使用的是以 x-api-key 请求头发送的 Crawlora API 密钥。这些 Crawlora 路由不需要单独的 Metaculus 密钥。
API 是否支持分类、锦标赛和项目页面？
在当前目录支持的范围内是支持的。Metaculus 端点集合包括带 slug 和 limit 参数的分类、锦标赛和项目问题列表路由。
API 是否返回结构化 JSON？
是的。Metaculus 端点返回规范化 JSON，并在可用时提供文档化的请求参数、响应架构、示例以及经过 Playground 测试的调用。
如何用 Crawlora 抓取 Metaculus？
先通过 Crawlora 的 /metaculus/questions 端点或分类、项目、锦标赛路由列出问题，再对 /metaculus/question/{id}、/metaculus/question/{id}/forecasts 和 /metaculus/question/{id}/forecast-history 发起 GET 请求，获取当前综合预测及其历史。所有调用都是针对公开页面的 GET 请求。
支持哪些综合预测方法？
recency_weighted、unweighted 和 single_aggregation，作为 method 参数传给 /metaculus/question/{id}/forecast-history 和 /metaculus/question/{id}/options 端点。
是否返回评论正文？
不会。comments-feed 和 top-comments 端点按评论顺序返回近期及高赞公开评论所对应的问题，但不包含上游的评论正文。
如果 Metaculus 返回登录页或验证页会怎样？
端点会直接失败：返回一个有文档说明的错误，而不是部分数据或已认证的数据，并且该次调用不会按成功计费。
这是预测建议吗？
不是。这些端点返回的是用于研究和跟踪的公开社区预测和元数据，不构成投资、预测或决策建议。
Guides
Metaculus guides and posts
Read Crawlora guides and comparisons that use the Metaculus API.
从 Metaculus API 开始
在 Playground 中测试 Metaculus questions，在 Docs 中查看端点架构，并在价格页面比较基于额度的用量。