Skip to content

Searching

The search command finds papers by keywords across every configured source, deduplicates the results, and returns a single merged list.

opencite search "transformer attention mechanism"

By default the search fans out to all available sources concurrently. Each source contributes its hits; OpenCite then merges duplicates (DOI match first, then fuzzy title match) and tracks which APIs contributed to each paper.

Choose your sources

Restrict the search to a single source with --source:

opencite search "AlphaFold protein" --source openalex
opencite search "EEG microstates" --source pubmed

Restricting sources is the fastest, most deterministic way to search, because OpenCite no longer waits on slower upstreams. It is the right choice when you know which database has the best coverage for your topic.

--source value Database
all (default) Every configured source
openalex OpenAlex
s2 Semantic Scholar
pubmed PubMed
arxiv arXiv
biorxiv bioRxiv
medrxiv medRxiv
osf OSF Preprints
zenodo Zenodo
figshare Figshare
crossref CrossRef
core CORE

Filter and sort

opencite search "tDCS motor cortex" \
    --max 30 \
    --year-from 2020 \
    --year-to 2024 \
    --oa-only \
    --sort citations
  • --max N: cap the number of results (applied after dedup and sort).
  • --year-from / --year-to: restrict to a publication-year window.
  • --oa-only: keep only open-access papers.
  • --sort relevance|citations|year: order the merged list. relevance is the default; citations surfaces the most influential work; year shows the newest first.

Canonical papers in a field

When you want the landmark papers rather than the most recent, use canonical. It searches and ranks by citation count, with a floor you control:

opencite canonical "deep learning" --min-citations 5000
opencite canonical "diffusion models" --max 10 --year-from 2020
  • --min-citations N: drop papers below this citation count.
  • --max N and --year-from work as they do for search.

Save results for later

Any search can be written to a file in your format of choice:

opencite search "spiking neural networks" --max 50 -f json -o snn.json
opencite search "spiking neural networks" --max 50 -f bibtex -o snn.bib

The JSON output is the natural input for batch PDF downloads. See Output Formats for the full list.