Find related papers
arxiv_related_papersFind papers similar to, cited by, or citing a known arXiv paper using Semantic Scholar, including non-arXiv results with authors, year, citations, and links.
Instructions
Find papers related to a known arXiv paper: similar, cited, or citing.
arXiv has no citation or similarity data, so this tool uses Semantic Scholar. Results can include papers that are not on arXiv (they have no arXiv ID and link to Semantic Scholar instead). Very new preprints may not be indexed yet.
Args: params (RelatedPapersInput): - paper_id (str): arXiv ID or URL, e.g. '1706.03762' - relation (str): similar | references | cited_by - max_results (int): Results per page (1 - 50, default 10) - start (int): Pagination offset (references and cited_by only) - response_format (str): markdown | json
Returns: str: List of related papers with authors, year, citation count, link and abstract snippet.
JSON schema when response_format='json':
{
"paper_id": str, "relation": str, "source": "semantic_scholar",
"start": int, "count": int, "has_more": bool,
"next_start": int | null,
"papers": [{ "title", "authors", "year", "published", "arxiv_id",
"doi", "venue", "citation_count", "url", "abstract" }]
}Examples: - Papers like Attention Is All You Need: paper_id='1706.03762' - What GPT-4's report builds on: paper_id='2303.08774', relation='references' - Follow-up work on Adam: paper_id='1412.6980', relation='cited_by'
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"params": "<params>"
},
"name": "arxiv_related_papers"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| params | Yes |
Output Schema
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |