Researcharr is a fork of Readarr refocused on academic papers and DOIs. It fetches metadata from multiple scholarly sources, searches research-first indexers, and automates grabs/renames via the familiar Servarr workflow.
- Metadata aggregator that merges CrossRef (authoritative DOI registry) with OpenAlex (concepts, institutions, citations).
- Research-friendly indexers: LibGen (known good mirrors), Unpaywall, ArXiv, Sci-Hub DOI search, PubMed Central.
- DOI-safe matching: download decisions backfill author/title only when DOI matches.
- Path safety fixes: long multi-author filenames are truncated to avoid filesystem errors.
- BibTeX export endpoint for any paper:
/api/v1/book/{id}/export/bibtex.
- Backend:
dotnet build src/Readarr.sln(or./build.shfor full pipeline). - Frontend dev:
yarn install --frozen-lockfile && yarn startfromfrontend/. - Run host:
dotnet run --project src/NzbDrone.Host/Readarr.Host.csproj --framework net6.0(defaults to port 7337). - Docker: build with
-p:RunAnalyzers=falseif StyleCop nags during publish.
- Preferred LibGen mirrors:
libgen.li,libgen.vg,libgen.la,libgen.bz,libgen.gl. - DOI lookups fetch merged metadata; title/author backfill is gated by DOI equality to avoid false positives.
- PubMed Central requests use XML and free full-text filters; Unpaywall authors are parsed from
raw_author_name.
- Expand metadata merge (better edition-level enrichment, citations).
- Additional academic sources (e.g., Semantic Scholar) and tests for indexer parsing.
- Optional Calibre-style tagging for imported papers.
- GNU GPL v3 (inherits from upstream Readarr).
