Arxiv Tools#

Configuration File: arxiv_tools.json Tool Type: Local Tools Count: 2

This page contains all tools defined in the arxiv_tools.json configuration file.

Available Tools#

ArXiv_get_pdf_snippets (Type: ArXivPDFSnippetsTool)#

Fetch an arXiv paper’s PDF and return bounded text snippets around provided terms. Uses markitdow…

ArXiv_get_pdf_snippets tool specification

Tool Information:

  • Name: ArXiv_get_pdf_snippets

  • Type: ArXivPDFSnippetsTool

  • Description: Fetch an arXiv paper’s PDF and return bounded text snippets around provided terms. Uses markitdown to convert PDF to markdown. All arXiv papers are freely available (no paywall). Useful when key details are in the full text rather than abstract.

Parameters:

  • arxiv_id (string) (optional) arXiv paper ID (e.g., ‘2301.12345’ or ‘arXiv:2301.12345’). The tool will build the PDF URL automatically.

  • pdf_url (string) (optional) Direct PDF URL (e.g., ‘https://arxiv.org/pdf/2301.12345.pdf’).

  • terms (array) (required) Terms to search for in the extracted full text (case-insensitive).

  • window_chars (integer) (optional) Context window size (characters) before and after each match.

  • max_snippets_per_term (integer) (optional) Maximum number of snippets returned per term.

  • max_total_chars (integer) (optional) Hard cap on total characters returned across all snippets.

Example Usage:

query = {
    "name": "ArXiv_get_pdf_snippets",
    "arguments": {
        "terms": ["item1", "item2"]
    }
}
result = tu.run(query)

ArXiv_search_papers (Type: ArXivTool)#

Search arXiv using the public Atom API with exact OpenSearch totals and recoverable start paginat…

ArXiv_search_papers tool specification

Tool Information:

  • Name: ArXiv_search_papers

  • Type: ArXivTool

  • Description: Search arXiv using the public Atom API with exact OpenSearch totals and recoverable start pagination. Begin with start=0 and continue with metadata.next_start while metadata.has_more is true. Date filters constrain submittedDate without replacing the numeric pagination offset.

Parameters:

  • query (string) (required) Search query for arXiv papers. Unquoted, unprefixed space-separated keywords are ANDed together (every keyword must match) – adding more keywords NARROWS results, it does not broaden them. To search multiple synonyms/phrases and broaden recall, join them with explicit OR, e.g. ‘“machine learning” OR “deep learning” OR “neural networks”’. Supports field prefixes (au:, ti:, abs:, cat:, all:) and AND/OR/ANDNOT boolean operators, which are passed through as-is.

  • limit (integer) (optional) Number of papers to return (1-200). This sets the maximum number of papers retrieved from arXiv.

  • start (integer) (optional) Zero-based result offset. For the next page, add the number of papers returned by the current call.

  • sort_by (string) (optional) Sort order for results. Options: ‘relevance’, ‘lastUpdatedDate’, ‘submittedDate’

  • sort_order (string) (optional) Sort direction. Options: ‘ascending’, ‘descending’

  • date_from (string) (optional) Filter results from this date (format: YYYY-MM-DD). Uses submittedDate, the date of the paper’s FIRST version – a paper revised/updated after this date but originally submitted earlier is still excluded.

  • date_to (string) (optional) Filter results up to this date (format: YYYY-MM-DD). Uses submittedDate, the date of the paper’s FIRST version – a paper revised/updated after this date but originally submitted earlier is still included.

Example Usage:

query = {
    "name": "ArXiv_search_papers",
    "arguments": {
        "query": "example_value"
    }
}
result = tu.run(query)