SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsacademic-search

academic-search

Data & analytics

View repository

Updated 10-15 16:47

About this skill

Design Philosophy

6 authoritative academic data sources, with Chinese and English dual-path parallelism:

用户 "高温热泵论文"
        │
   ├──→ 🔬 Semantic Scholar     → 英文论文 + DOI + 引用次数
   │
   └──→ 🇨🇳 秘塔搜索            → 中文学术 + 产业报告
                 │
                 ▼
            合并呈现：双语学术搜索

⚡ First-time use: configure API Keys (for the best experience)

When a user triggers an academic search for the first time after installing this Skill, execute the following process:

Step 1: Guide configuration (can be skipped)

"Academic search is ready — Chinese and English dual-path parallel, free and ready out of the box.

If you want a better experience, you can configure Keys (optional, can be added anytime):

  1. Semantic Scholar Key: removes the 100 requests/5 minutes rate limit → high speed and stable
  2. Meta Academic Key: upgrades Chinese search quality (broader and deeper)
  3. OpenAlex Key: unlocks the largest open academic library (250 million+, all disciplines)

Do you want to configure now or use it as is for now? Just tell me."

Step 2: Write to .env

User reply Action
"Have S2 Key: xxx" Write SEMANTIC_SCHOLAR_API_KEY=xxx
"Have Meta Key: xxx" Write METASO_API_KEY=xxx
"Have both" Write both
"Have neither" Use the S2 free tier; Meta is temporarily unavailable (can be configured later anytime)

Step 3: Confirm

✅ 配置完成！

Semantic Scholar: [配 Key ✓  高速模式] / [免费层 100req/5min]
秘塔搜索:       [配 Key ✓  中文可用] / [未配置 ✗  仅英文（可后期补配）]

后期补配 Key：编辑 <skill_dir>/.env

Rate Comparison

Source No Key With Key
Semantic Scholar 100 req/5 min (may be rate limited) High speed, stable
Meta Search ❌ Unavailable ✅ Real-time Chinese academic search
Crossref ✅ Free and unlimited —
PubMed ✅ 3req/s 10req/s
arXiv ✅ Free and unlimited —
Zenodo ✅ Free and unlimited —

🌐 Dual-path search (default behavior)

Every academic search runs Chinese and English dual-path in parallel — this is the default behavior, not an option.

用户 "高温热泵论文"
        │
   ├──→ 🔬 Semantic Scholar     → 英文论文 + DOI + 引用次数
   │
   └──→ 🇨🇳 Exa + 中文学术站点   → 免费中文论文（知网/万方/维普/百度学术）
                 │
                 ▼
            合并呈现：双语学术搜索

Execution Rules

Scenario Behavior
User searches any academic keyword Simultaneouslycall the English path + Chinese path, returning in parallel
User explicitly says "English papers" Call only the English path
User explicitly says "Chinese papers" Call only the Chinese path

Fallback Strategy

                    ┌── 英文路 ──────────────────────┐
                    │                                  │
用户搜 "关键词" ───┤   1. Semantic Scholar（主）       │
                    │      ↓ 限流/故障                 │
                    │   2. OpenAlex（降级）             │
                    │      ↓ 也挂了                     │
                    │   3. 仅返回中文结果 + 报错提示     │
                    │                                  │
                    └── 中文路 ──────────────────────┘
                           │
                     有秘塔Key？
                      ├── 是 → 秘塔搜索（主）
                      │         ↓ 故障
                      │       papers_zh（降级）
                      │
                      └── 否 → papers_zh（多后端自动降级）
                                │
                                ├── Exa（主）
                                │     ↓ 不可用
                                ├── Brave Search
                                │     ↓ 不可用
                                ├── web_search
                                │     ↓ 不可用
                                └── OpenAlex(lang:zh)（兜底）
                                      ↓ 不可用
                                     仅返回英文结果 + 提示
Fallback path Trigger condition Effect
S2 → OpenAlex S2 429 rate limit / timeout Use OpenAlex to continue searching in English
Meta → papers_zh Meta API failure Fall back to free Chinese search
papers_zh → skip Chinese Exa unavailable English-only results + prompt "configure Meta Key to unlock Chinese"
All unavailable → error Rare Prompt the user to check the network

Output Format

## 🔬 英文学术论文 (Semantic Scholar, N 条)
### 1. [标题]
- 期刊 / 引用次数 / DOI ...

## 🇨🇳 中文学术结果 (papers_zh, N 条)
### 1. [标题]
- 来源 / 摘要 / 链接 ...

Trigger Conditions

Activate when a user query involves the following scenarios:

  • "find papers", "search literature", "what related research exists", "latest papers"
  • "DOI lookup", "information about this paper", "citation count"
  • "search PubMed", "biomedical literature", "clinical trials"
  • "arXiv preprints", "latest preprint"
  • "research datasets", "dataset"

8 major search sources

Source Backend Key Purpose
papers Semantic Scholar Recommended English academic papers (200M+)
openalex OpenAlex Optional Largest open academic library (250M+, all disciplines)
papers_zh Exa + Chinese academic sites Not required Free Chinese papers (CNKI/Wanfang/VIP/Baidu Scholar)
citation Crossref Not required DOI metadata + citation relationships
biomedical PubMed + Europe PMC Not required Biomedical literature
preprint arXiv Not required Latest preprints
dataset Zenodo Not required Research datasets
metaso Metaso Search Recommended Real-time Chinese academic/technical search

See the full API documentation at references/apis.md。

Usage

Semantic paper search (Semantic Scholar)

GET https://api.semanticscholar.org/graph/v1/paper/search?query={url_encoded_query}&limit=10&fields=title,authors,year,citationCount,journal,externalIds

Use webfetch：

webfetch "https://api.semanticscholar.org/graph/v1/paper/search?query=solid+state+battery+electrolyte&limit=5&fields=title,authors,year,citationCount,journal,externalIds"

Use curl：

curl "https://api.semanticscholar.org/graph/v1/paper/search?query=solid+state+battery&limit=5&fields=title,authors,year,citationCount,journal,externalIds"

Rate limit: 100req/5min (free), higher with a Key

Free Chinese paper search (papers_zh)

Automatic backend fallback:Exa → Brave → web_search → OpenAlex(lang:zh)

site:cnki.net OR site:cqvip.com OR site:wanfangdata.com.cn OR site:xueshu.baidu.com OR site:arxiv.org

Examples for each backend：

# Exa（最佳）
web_search_exa "高温热泵 论文 site:cnki.net OR site:cqvip.com OR site:wanfangdata.com.cn"

# Brave Search
brave_web_search "高温热泵 论文 site:cnki.net OR site:cqvip.com"

# 通用搜索
web_search "高温热泵 论文 site:cnki.net OR site:cqvip.com"

# OpenAlex（最终兜底，中文相关性弱）
curl "https://api.openalex.org/works?search=高温热泵&filter=language:zh"

Covers CNKI, Wanfang, VIP, Baidu Scholar, and arXiv. Does not depend on any specific tool.

Precise DOI lookup (Crossref)

curl "https://api.crossref.org/works/10.1038/s41586-021-03819-2"

Returns complete metadata: title, authors, journal, date, abstract, reference list, citation count.

Biomedical search (PubMed)

# Step 1: 搜索
curl "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=mRNA+vaccine+cancer&retmax=5&retmode=json"
# Step 2: 获取详情
curl "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esummary.fcgi?db=pubmed&id=<PMID>&retmode=json"

Biomedical search (Europe PMC, simpler)

curl "https://www.ebi.ac.uk/europepmc/webservices/rest/search?query=COVID-19+vaccine&resultType=core&pageSize=5&format=json"

Preprint search (arXiv)

curl "http://export.arxiv.org/api/query?search_query=all:large+language+model+reasoning&start=0&max_results=5&sortBy=submittedDate&sortOrder=descending"

Metaso Search (Chinese academic/technical)

curl -X POST "https://metaso.cn/api/v1/search" \
  -H "Authorization: Bearer $METASO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"q":"固态电池 最新进展","n":5}'

Covers real-time content such as Chinese academic papers, industry reports, technical blogs, and Bilibili videos.

Output format

Results are presented in a structured way:

## [数据源] 搜索结果 (N 条)

### 1. [标题]
- 作者: [作者列表]
- 期刊/年份: [Venue] ([Year])
- DOI: [DOI]
- 引用次数: [N]（仅 Semantic Scholar）
- 链接: [URL]

Example conversation

User: "What are the latest papers on solid-state batteries?"

Agent executes:

curl "https://api.semanticscholar.org/graph/v1/paper/search?query=solid+state+battery+electrolyte&limit=10&fields=title,authors,year,citationCount,journal,externalIds"

User: "Information about this paper? DOI: 10.1038/s41586-021-03819-2"

Agent executes:

curl "https://api.crossref.org/works/10.1038/s41586-021-03819-2"

User: "Search PubMed for mRNA cancer vaccines"

Agent executes:

curl "https://eutils.ncbi.nlm.nih.gov/entrez/eutils/esearch.fcgi?db=pubmed&term=mRNA+vaccine+cancer&retmax=10&retmode=json"

User: "What new preprints are there on arXiv about MoE architectures?"

Agent executes:

curl "http://export.arxiv.org/api/query?search_query=all:mixture+of+experts+language+model&start=0&max_results=5"

PDF full-text download

Supports downloading Open Access paper PDFs:

Source PDF support Description
papers (S2) ✅ openAccessPdf field, automatically retrieved for OA papers
preprint (arXiv) ✅ arxiv.org/pdf/{id}.pdf, all free
biomedical (PubMed) ⚠️ Only PMC free full text
Other sources ❌ Metadata only, no PDF

CLI download：

python academic_search.py papers "solid state battery" --limit 3 --pdf -o ./downloads
python academic_search.py preprint "transformer attention" --pdf -o ./downloads

Agent usage: When search results contain pdf_url or arxiv_id , use webfetch / curl to download.

Extension guide

Adding a new API takes only two steps:

  1. In references/apis.md Add API documentation
  2. Add the corresponding row in the document subdomain table in this document