domains / research / openalex-academic-search / openalex

openalex

Skill card

combined

Both copies target OpenAlex, but they prescribe different tools and workflows. The first copy is a Python fetcher skill (openalex_fetch.py) with a strict resolution chain, optional .env API key, and filters like min-citations and type. The second copy is a Node CLI skill (openalex-skill) with entity types, ID formats, cited-by/references/related, autocomplete, group-by, PDF download, and output formats (summary/detail/json/bibtex). Neither is a reproduction of the other; crowning one would hide a real alternative implementation and a set of capabilities the other lacks.

Repositories

Other copies

Combined skill

# OpenAlex Academic Search

Search academic papers, authors, institutions, and citations via the OpenAlex API.

## When to Use

Use this skill when the user wants to:
- Search for academic papers or scholarly works
- Find author, institution, or journal information
- Track citations (cited-by, references, related)
- Look up metadata by DOI, ORCID, or OpenAlex ID
- Analyze publication trends or research metrics
- Download open-access PDFs

## Setup

### Option A: Python Fetcher (wanshuiyin style)

1. Install Python 3.7+ and requests:
   ```bash
   pip install requests
   ```
2. Optional: Create `.claude/.env` with `OPENALEX_API_KEY` and `OPENALEX_EMAIL`.
3. Resolve `openalex_fetch.py` via the canonical chain (`.aris/tools/`, `tools/`, `$ARIS_REPO/tools/`).

### Option B: Node CLI (brycewang-stanford style)

1. Install the CLI:
   ```bash
   npm install -g openalex-skill
   openalex --help
   ```
2. Configure API key per `references/setup.md` if needed.

## Core Commands

### Search Works

**Python fetcher:**
```bash
python3 "$OPENALEX_FETCHER" search "QUERY" --max 10
```

**Node CLI:**
```bash
openalex works search "QUERY" --per-page 5
```

### Filters and Sort

**Python fetcher overrides (append to arguments):**
- `— max: 20` — return up to 20 results
- `— year: 2023-` — papers from 2023 onward
- `— year: 2020-2023` — papers from 2020 to 2023
- `— type: article` — only journal articles
- `— type: preprint` — only preprints
- `— open-access` — only open access papers
- `— min-citations: 50` — minimum 50 citations
- `— sort: citations` — sort by citation count
- `— sort: date` — sort by publication date

**Node CLI filters:**
```bash
openalex works list \
  --filter publication_year:2024 \
  --filter is_oa:true \
  --sort cited_by_count:desc \
  --per-page 10
```

### Get Specific Work

**Python fetcher:**
```bash
python3 "$OPENALEX_FETCHER" work "10.1109/TWC.2024.1234567"
```

**Node CLI:**
```bash
openalex works get W2741809807
openalex works get https://doi.org/10.1038/nature12373
openalex works get 10.1038/nature12373
```

### Citation Tracking (Node CLI only)

```bash
# Papers that cite this work
openalex works cited-by W2741809807 --per-page 5

# Papers this work references
openalex works references W2741809807 --per-page 5

# Related works
openalex works related W2741809807 --per-page 5
```

### Author and Institution Lookup (Node CLI only)

```bash
openalex authors search "Author Name" --per-page 3
openalex authors get https://orcid.org/0000-0002-3141-5845
openalex institutions autocomplete "tsinghua"
```

### Group By (Node CLI only)

```bash
openalex works group --by publication_year \
  --filter author.id:A5070829652
```

### Download PDF (Node CLI only)

```bash
openalex works download https://doi.org/10.48550/arXiv.1706.03762
openalex works download W2741809807 -o paper.pdf
```

## Output Formats (Node CLI)

- `summary` (default) — Concise one-line format
- `detail` — Human-readable with inline lists
- `json` — Full structured payload
- `bibtex` — BibTeX entries
- `--field <path>` — Client-side projection
- `--select <field>` — Server-side selection

## Conflicts

- **wanshuiyin/Auto-claude-code-research-in-sleep** uses a Python helper script with a strict resolution chain and requires `pip install requests`; **brycewang-stanford/Auto-Empirical-Research-Skills** uses a Node CLI installed via `npm` and does not require Python.
- **wanshuiyin/Auto-claude-code-research-in-sleep** supports `min-citations` and work-type filters; **brycewang-stanford/Auto-Empirical-Research-Skills** does not document these but supports entity types, autocomplete, group-by, and PDF download.
- **wanshuiyin/Auto-claude-code-research-in-sleep** uses `.claude/.env` for API key; **brycewang-stanford/Auto-Empirical-Research-Skills** references a separate setup document.
- **wanshuiyin/Auto-claude-code-research-in-sleep** has a fixed `MAX_RESULTS` default of 10; **brycewang-stanford/Auto-Empirical-Research-Skills** uses `--per-page` with no fixed default in the skill text.