hotdaily

When the user asks about technology trends, startup news, or wants to learn about the latest technology articles, use this skill to fetch data from the HotDaily API.

Core Capabilities

HotDaily is an AI-driven curated daily digest of technology and startups. Every day it selects content from communities such as Hacker News and Lobsters, and after AI reading and value assessment, presents it to readers.

Available Endpoints

Digest-related

Today's Digest
GET https://api.hotdaily.top/v1/digests/today

Returns the latest issue of today's digest (morning/noon/evening), including all articles the AI has read on behalf of readers, value judgments (deep read/skim), Chinese summaries, and reading time.

Digest for a Specified Date
GET https://api.hotdaily.top/v1/digests/{date}

Fetches the digest for a historical day. Date format:YYYY-MM-DD, for example 2026-06-22。

Digest Archive List
GET https://api.hotdaily.top/v1/digests

Returns the date list of all historical digests, sorted in reverse order.

More Selections (Overflow Page)
GET https://api.hotdaily.top/v1/more?page={N}

Good articles that met the criteria that day but were not selected for the top 30, fetched with pagination.page Defaults to 1.

Trend-related

Today's Trends
GET https://api.hotdaily.top/v1/trends/today

Returns the day's technology trend signals, intensity curves, and evidence items. Each trend includes: title, summary, why it matters, status (new/heating/sustained/cooling), confidence, intensity score, and evidence sources.

Weekly/Monthly Report
GET https://api.hotdaily.top/v1/trends?window={window}&date={date}

• window: today | week | month
• date: anchor date (YYYY-MM-DD), optional, defaults to the latest

The weekly report is generated every Monday (aggregating signals from the past 7 days), and the monthly report is generated on the 1st of every month (aggregating signals from the past 30 days).

Trend Archive List
GET https://api.hotdaily.top/v1/trends/archive

Returns the date and window list of all trend reports (daily/weekly/monthly).

Item Details

Single Article Details
GET https://api.hotdaily.top/v1/items/{id}

Fetches the complete information of a single article: title, URL, value rationale, full summary, review information, reading time, and body (if renderable).

Common Return Fields

Common Fields of the Digest API

{
  "date": "2026-06-22",
  "edition": "晚报",
  "serverDate": "2026-06-22",
  "readCount": 156,
  "selectedCount": 30,
  "qualifiedCount": 45,
  "items": [
    {
      "id": "0a831d51-33eb-4963-931c-240d82ceba84",
      "title": "Article Title in English",
      "url": "https://example.com/article",
      "valueLight": "green",
      "reason": "为什么值得读的中文理由",
      "valueVerdict": "一句中文判词",
      "summaryZh": "中文摘要",
      "readingMinutes": 5,
      "source": "hacker-news",
      "signals": {
        "points": 342,
        "numComments": 78
      },
      "typeTag": "research"
    }
  ]
}

Field Description：

• valueLight: green = deep read (in-depth value),yellow = skim (quick glance)
• readCount: total number of articles the AI has read on behalf of readers
• selectedCount: number of digest selections (default 30 articles)
• qualifiedCount: total number meeting the criteria (including those overflowing to /more)
• reason: Chinese value rationale
• valueVerdict: a one-sentence Chinese verdict for readers, may be null
• summaryZh: a one-sentence Chinese summary, may be null

Common Fields of the Trend API

{
  "window": "today",
  "anchorDate": "2026-06-22",
  "generatedAt": 1782072932123,
  "trends": [
    {
      "id": "trend:today:abc",
      "title": "AI 模型路由成为基础设施新热点",
      "summary": "多家公司推出模型路由解决方案，自动选择最优模型处理请求",
      "whyItMatters": "降低成本同时提升响应质量，成为 AI 应用的关键中间层",
      "status": "heating",
      "score": 91,
      "confidence": "high",
      "sparkline": [3, 5, 7, 9],
      "evidence": [
        {
          "itemId": "...",
          "title": "OpenRouter launches new routing algorithm",
          "source": "hacker-news",
          "rank": 1
        }
      ]
    }
  ]
}

Field Description：

• status: new(newly appeared) / heating(heating up) / sustained(sustained) / cooling(cooling down)
• confidence: high / medium / low
• score: trend intensity score
• sparkline: array of intensity curve values, used to visualize changes in trend intensity

Common Fields of the Item Details API

{
  "id": "...",
  "title": "Article Title",
  "url": "https://...",
  "valueLight": "green",
  "valueReason": "价值理由",
  "summaryTriageZh": "简短摘要",
  "summaryDetailZh": "详细摘要",
  "readingMinutes": 5,
  "bodyKind": "full",
  "content": {
    "markdown": "# 正文\n\n...",
    "charCount": 5234
  },
  "review": {
    "quality": "ok",
    "compliance": "pass",
    "note": ""
  }
}

bodyKind Description：

• full: full body is renderable
• native: native content (forum posts, etc.)
• pdf: PDF document (already read multimodally)
• abstract: academic paper abstract
• summary: summary only
• none: no body

How users understand it:

• The list page mainly shows:title、valueLight、reason、valueVerdict、summaryZh
• The trend page mainly shows:title、summary、whyItMatters、status、evidence
• The detail page mainly shows:valueReason、summaryTriageZh、summaryDetailZh、content

Usage Patterns

1. User asks "what's worth reading today"

curl -s https://api.hotdaily.top/v1/digests/today

Presentation Suggestions：

• Group by valueLight (deep read vs skim)
• Each article includes: title, reason, reading time, link
• Prompt: "AI has read {readCount} articles for you, selected {selectedCount}"

2. User asks "recent technology trends"

curl -s https://api.hotdaily.top/v1/trends/today

Present suggestions：

• Each trend includes: title, summary, why it matters
• Mark status (🆕 new / 🔥 heating up / 🔄 sustained / ❄️ cooling down)
• List main evidence sources (2-3 representative items)

3. User asks "how is AI infrastructure doing recently"

curl -s "https://api.hotdaily.top/v1/trends?window=week"

By title, summary,whyItMatters and evidence items, determine which trends belong to the AI / infrastructure direction, then summarize the direction of evolution.

4. User asks "show me the June 20 daily report"

curl -s https://api.hotdaily.top/v1/digests/2026-06-20

5. User wants details of a certain article

curl -s https://api.hotdaily.top/v1/items/{id}

Show the full summary, value analysis, original link; if bodyKind is full or native, you can show excerpts from the main text.

Call examples (for those using the interface)

If a user or agent wants to call the public interface directly, they can read it like this:

# 获取今日日报
curl -s https://api.hotdaily.top/v1/digests/today | jq '.items[] | {title, valueLight, reason}'

# 获取今日趋势
curl -s https://api.hotdaily.top/v1/trends/today | jq '.trends[] | {title, summary, status}'

# 获取周报
curl -s "https://api.hotdaily.top/v1/trends?window=week" | jq '.trends[0]'

Usage notes

• The API requires no authentication, direct GET access
• Today's daily report and trends are not cached, updated in real time
• Historical daily reports have edge caching, fast response
• After parsing JSON, organize and present it according to the user's intent, do not directly forward the raw JSON
• Trend generation timing: automatic daily, weekly report generated on Monday, monthly report generated on the 1st of each month
• If the API returns 404, it means the report for that date has not yet been published or the trend has not yet been generated
• Content language: titles are the original English titles, summaries, reasons, and trend analyses are all in Chinese

Typical usage

User: What is worth reading today?

Assistant behavior：

1. Call /v1/digests/today
2. Classify by value light (deep read / skim)
3. Each article includes: title (English), reason (Chinese), duration, link
4. Header summary: "AI has read 156 articles for you, selected 30"

---

User: What trends are there in the AI field recently?

Assistant behavior：

1. Call /v1/trends/today
2. Determine whether it belongs to the AI field by title, summary, and evidence
3. Each trend includes: title, summary, why it matters, status (🔥/🔄/❄️) If you need to attach Chinese labels, it is recommended to map them as:new=新出现、heating=升温、sustained=持续、cooling=降温
4. Optional: list 2-3 representative evidence items

---

User: Show me this week's technology trends report

Assistant behavior：

1. Call /v1/trends?window=week
2. Show all trends, explain them in the order returned by the interface (usually already sorted by strength)
3. Each trend notes its direction of evolution (new/heating/sustained/cooling)

---

User: Talk about this article in detail [provide ID]

Assistant behavior：

1. Call /v1/items/{id}
2. Show the full summary (summaryDetailZh）
3. value reason (valueReason）
4. If there is a main text (bodyKind is full), you can excerpt key paragraphs
5. Give the original link

Presentation principles

• Prioritize giving the user a Chinese summary and Chinese reasons for its value; do not just throw raw JSON at them
• For titles, usually keep the original English title so the user can search for the original text
• If the user asks "what are the recent trends," first look at today
• If the user asks for "this week/this month," then switch to week / month
• If the user wants to dig deeper into a specific article, then call /v1/items/{id}