Paper Fetch

A practical wrapper around the academic paper fetching gauntlet. It takes a DOI, arXiv ID, or paper title and walks through Unpaywall, Semantic Scholar, arXiv, PMC, and preprint servers in order until something returns a PDF. The title resolver hits Crossref first, then falls back to Semantic Scholar when confidence is low or the paper is arXiv-only. Batch mode handles lists of DOIs. There's optional institutional access (tries publisher direct links if you're on a university network) and a Sci-Hub fallback that's on by default but easy to disable. The JSON output contract is thorough, complete with request IDs, sources tried, and partial success states for batch jobs. It's the kind of thing you write once you've manually chased down enough papers through broken links.