Rebiber is a bibliography management tool that replaces unofficial or arXiv BibTeX entries with official records from DBLP or the ACL Anthology.
The tool solves the problem of maintaining accurate and consistent bibliographic data in research papers. When researchers cite preprints or unofficial versions, their bibliography entries lack standardized formatting and complete metadata. Rebiber addresses this by matching BibTeX entries against DBLP and ACL Anthology databases, automatically replacing incomplete or informal entries with their official counterparts while preserving the original cite keys. The matching process checks author names and titles to ensure accuracy, with an optional live-lookup feature that searches DBLP by title when local databases miss an entry.
Rebiber suits researchers and teams managing LaTeX documents with BibTeX bibliographies, particularly those working in machine learning and natural language processing where DBLP and ACL Anthology coverage is strong. The tool works offline by default using packaged database dumps, making it suitable for reproducible workflows. It offers fine-grained control through options to shorten venue names, remove specific fields, protect acronyms in titles, sort entries, and filter replacements to only cited references. A dry-run mode lets users review changes before applying them. The tool requires Python 3.10 or newer and must be installed from GitHub rather than PyPI.
Development activity shows consistent maintenance with recent updates addressing core dependencies. The project migrated to bibtexparser 2 and raised the minimum Python version requirement, indicating active alignment with upstream library changes. Formatting improvements preserve specific whitespace conventions in BibTeX files, suggesting attention to preserving user intent during transformations. The tool maintains packaged database dumps that can be refreshed from the repository's main branch, indicating ongoing curation of reference data.
