Jabkit

jabkit — JabRef's CLI Toolkit

jabkit exposes JabRef's BibTeX/biblatex handling as a command-line application.

Installation

Preferred: the native binary — a single self-contained executable, no JDK or JBang needed, instant startup. Download and unpack for your platform:

# Linux (amd64)
curl -fL https://builds.jabref.org/main/linux-amd64/tools/jabkit-native_linux.tar.gz | tar xz
./jabkit/jabkit --help

# Linux (arm64)
curl -fL https://builds.jabref.org/main/linux-arm/tools/jabkit-native_linux_arm64.tar.gz | tar xz
./jabkit/jabkit --help

# macOS (Apple Silicon)
curl -fLO https://builds.jabref.org/main/macOS-silicon/tools/jabkit-native_macos-silicon.zip
unzip jabkit-native_macos-silicon.zip
.jabkit/jabkit --help

On other platforms (e.g. Windows, Intel macOS), run jabkit via JBang instead — no JDK setup needed. Install JBang if missing:

curl -Ls https://sh.jbang.dev | bash -s - app setup   # Linux/macOS

iex "& { $(iwr -useb https://ps.jbang.dev) } app setup"   # Windows

On linux-x64, mac-aarch64 or windows-x64, setting the environment variable JBANG_USE_NATIVE to true before the command above installs JBang's native binary instead of its JAR: one JDK download less and a faster start.

export JBANG_USE_NATIVE=true          # Bash/zsh, persist in ~/.bashrc or ~/.zshrc

$env:JBANG_USE_NATIVE = 'true'        # PowerShell, persist in $PROFILE

Persist it rather than setting it only for the install — the launcher re-reads it on every run and falls back to the JAR when it is unset. Other platforms have no native build and the install would fail, so leave it unset there.

Then install jabkit on the PATH (the same command also updates it):

jbang app install --fresh --force jabkit@jabref

One-off run without installation:

jbang --fresh jabkit@jabref --help

Global flags

Place before the subcommand:

• -p, --porcelain — script-friendly output (no banners/progress)
• -d, --debug — debug logging
• -v, --version — version info

Commands

Input files are passed positionally or via --input; both forms are equivalent. --input-format "*" auto-detects the input format.

Examples

# DOI to BibTeX
jabkit doi-to-bibtex 10.1145/3149935.3149942

# Fetch from an online catalogue into a file
jabkit fetch --provider ArXiv --query "test driven development" --output tdd.bib

# PDF paper to BibTeX entry (see the pdf-to-bibtex skill for details)
jabkit convert --input paper.pdf --input-format pdfMerged --output paper.bib

# Validate in CI (errorformat output, non-zero exit on findings)
jabkit -p check library.bib --output-format github-actions

# Citation keys with a custom pattern
jabkit citationkeys generate library.bib --pattern "[auth][year]" --output library.bib

# Cited works from a paper's "References" section (see the pdf-extract-references skill)
jabkit -p pdf extract-references paper.pdf

# Write BibTeX + XMP metadata into the PDFs linked from an entry
jabkit pdf update --citation-key Smith2020 --input library.bib --input-format bibtex

# Library subset actually cited in a LaTeX document
jabkit generate-bib-from-aux --aux paper.aux --input full-library.bib --output paper.bib

Notes for agents

• Always use -p/--porcelain when parsing output programmatically.
• fetch --provider matches JabRef's web-search fetcher names case-insensitively; on an unknown name, jabkit reports Could not find fetcher.
• URLs are accepted as input: jabkit convert --input https://example.org/refs.ris --input-format ris.