Build repeatable open-access full-text mining pipelines (discovery, retrieval, parsing, extraction, and QA) and feed structured outputs back into Zotero-centered evidence workflows. Use this whenever a user asks to mine many papers for entities, outcomes, or section-level data.
Scale from single-paper reading to reproducible corpus-level extraction while staying inside legal and licensing boundaries.
PMC, Europe PMC, CORE, arXiv).DOI, PMCID, arXiv ID).{
"paper_id": "doi|pmcid|arxiv",
"source": "pmc|core|arxiv|other",
"license": "string or null",
"sections": [
{"title": "Methods", "text": "..."}
],
"entities": [
{"text": "string", "type": "string"}
],
"numeric_findings": ["string"],
"provenance": {
"retrieved_at": "ISO-8601",
"source_url": "string"
}
}
retrieved, parsed, qa-passed, needs-review).