Google Scholar Scraper: Papers, Authors, Citations, BibTeX
Pricing
Pay per usage
Search Google Scholar at scale. Pulls paper metadata, author affiliations, h-index, cited by counts, citing paper lists, BibTeX, and PDF links. One row per paper. Pay per row.
Free text queries. Supports Scholar operators: "exact phrase", author:Hinton, intitle:transformer. Example: ["large language model", "author:LeCun convolutional"].
[]Scholar cluster IDs to fetch all versions of one paper, or its citing papers. Example: 17784817748666649498.
[]Scholar result URLs to enrich with full metadata (BibTeX, citing papers, related). Useful when you already have a list of papers and need citations.
[]Lower bound on publication year. 0 means no lower bound.
Upper bound on publication year. 0 means no upper bound.
Relevance is Scholar's default. Date sorts newest first.
Scholar interface language. Affects venue parsing and citation labels.
Include patent results. Off by default to keep results to peer reviewed papers.
Include legal case law results. Off by default.
For every paper with at least one author profile linked, fetch the author's affiliation, h-index, i10-index, and total citations. Adds one extra request per unique author.
For papers with N or more citations, fetch the list of papers that cite this work. Adds citingPapers[] to the row.
Only fetch citing papers when the paper has at least this many citations. Avoids wasting requests on low cited papers.
Cap on how many citing papers to fetch per source paper. Scholar returns 10 per page.
Pull the BibTeX export for every paper. Adds bibtex (string) to the row. Adds one extra request per paper.
For each paper, fetch every Scholar cluster version (preprint, published, repository copies). Adds versions[] to the row.
Hard cap on paper rows pushed per run. 0 means unlimited (Scholar caps at ~1000 results per query).
Pages of 10 results to walk per query. Scholar caps at 100 pages (1000 results) per query and aggressively rate limits past page 30.
Skip cluster IDs already pushed in previous runs. Keyed on Scholar cluster ID. Turn off to refresh stale rows.
Parallel browser pages. Keep at 1 to 2 unless you have a residential proxy pool.