doi2bib

Small CLI tools to work with bibtex entries: get entries from DOI/arXiv/PubMed IDs and format bibtex files.

Just so you know, there is now Zotero BIB on the browser that can do this (and more). I will keep maintaining doi2bib though, since it is an integral part of my workflow.

Tools

This package provides three CLI tools:

1. doi2bib - Get bibtex entries from DOI, arXiv ID, or PubMed ID (pretty printed with bibfmt)
2. bibfmt - Pretty print and format bibtex files (using very few dependencies)
3. bibdedup - Deduplicate BibTeX entries across multiple files

doi2bib Usage

$ doi2bib --help=plain
NAME
doi2bib - A little CLI tool to get the bibtex entries for DOIs, arXiv
IDs, or PubMed IDs.
SYNOPSIS
doi2bib [--input=FILE] [--output=OUTPUT] [OPTION]… [ID]…
DESCRIPTION
doi2bib fetches BibTeX entries for given DOI, arXiv, or PubMed IDs.
The tool automatically infers the type of identifier. You can force the
lookup of a DOI by using the form 'doi:ID' or an arXiv ID by using the
form 'arXiv:ID'. PubMed IDs always start with 'PMC'.
ARGUMENTS
ID A DOI, an arXiv ID or a PubMed ID. The tool tries to automatically
infer what kind of ID you are using. You can force the cli to
lookup a DOI by using the form 'doi:ID' or an arXiv ID by using
the form 'arXiv:ID'. PubMed IDs always start with 'PMC'. Multiple
IDs can be provided and will be processed in order.
OPTIONS
-i FILE, --input=FILE
The file is read line by line and processed sequentially, treating
each line as DOIs, arXiv IDs or PubMedIDs. Errors will be printed
on standard error but will not terminate the operation.
-o OUTPUT, --output=OUTPUT (absent=stdout)
Append the bibtex output to the specified file. It will create the
file if it does not exist.
COMMON OPTIONS
--help[=FMT] (default=auto)
Show this help in format FMT. The value FMT must be one of auto,
pager, groff or plain. With auto, the format is pager or plain
whenever the TERM env var is dumb or undefined.
--version
Show version information.
EXAMPLES
Process a single DOI:
$ doi2bib 10.1007/s10569-019-9946-9
Process multiple IDs:
$ doi2bib 1902.00436 arXiv:1609.01724 PMC2883744
Save bibtex entry to a file:
$ doi2bib doi:10.4171/JST/226 -o bibliography.bib
Process a file containing identifiers:
$ doi2bib --input dois.txt -o bibliography.bib
Read from stdin:
$ echo '10.1145/3357713.3384296' | doi2bib --input -
EXIT STATUS
doi2bib exits with:
0 on success.
123 on indiscriminate errors reported on standard error.
124 on command line parsing errors.
125 on unexpected internal errors (bugs).
BUGS
Report bugs to https://github.com/mseri/doi2bib/issues

The tool retrieves the bibtex entry using published details when possible.

bibfmt Usage

$ bibfmt --help=plain
NAME
bibfmt - A little CLI tool to pretty print bibtex files.
SYNOPSIS
bibfmt [OPTION]... [FILES]...
DESCRIPTION
bibfmt reads one or more BibTeX files, parses them, and outputs
formatted BibTeX entries.
Use '-' as a filename to read from stdin.
ARGUMENTS
FILES BibTeX files to format. Use '-' to read from stdin. Multiple
files can be specified and will be combined.
OPTIONS
-f, --force
Force mode: ignore parsing errors and output only successfully
parsed entries.
-l, --single-line
Force field values onto a single line by replacing newlines with a
space.
-o OUTPUT, --output=OUTPUT (absent=stdout)
Saves the pretty printed bib to the specified file. If not
specified, writes to stdout.
-q, --quiet
Quiet mode: suppress all output except errors.
-s, --strict
Enable strict parsing mode that rejects BibTeX files with
duplicate fields.
-v, --verbose
Enable verbose output showing which files are being read.
COMMON OPTIONS
--help[=FMT] (default=auto)
Show this help in format FMT. The value FMT must be one of auto,
pager, groff or plain. With auto, the format is pager or plain
whenever the TERM env var is dumb or undefined.
--version
Show version information.
EXAMPLES
Format a single file:
$ bibfmt bibliography.bib -o formatted.bib
Format multiple files:
$ bibfmt file1.bib file2.bib -o combined.bib
Read from stdin:
$ cat input.bib | bibfmt -
Combine stdin with files:
$ echo '@article{...}' | bibfmt - existing.bib -o output.bib
Format with strict mode:
$ bibfmt --strict bibliography.bib
EXIT STATUS
bibfmt exits with:
0 on success.
123 on indiscriminate errors reported on standard error.
124 on command line parsing errors.
125 on unexpected internal errors (bugs).
BUGS
Report bugs to https://github.com/mseri/doi2bib/issues

bibdedup Usage

$ bibdedup --help=plain
NAME
bibdedup - Deduplicate BibTeX entries across multiple files.
SYNOPSIS
bibdedup [OPTION]... FILES...
DESCRIPTION
bibdedup reads one or more BibTeX files, combines all entries, and
removes duplicates based on specified key fields.
By default, entries are considered duplicates if they have the same
title, author, and year (after whitespace normalization and
case-insensitive comparison).
ARGUMENTS
FILES BibTeX files to deduplicate. Use '-' to read from stdin.
Multiple files can be specified.
OPTIONS
-i, --interactive
Enable interactive mode to resolve conflicts. If not set,
automatically keeps the first occurrence of conflicting fields.
-k KEYS, --keys=KEYS
Comma-separated list of field names to use for duplicate
detection. Special key 'citekey' matches on citation keys.
Default: title,author,year
-o OUTPUT, --output=OUTPUT (absent=stdout)
Output file for deduplicated BibTeX. If not specified, writes to
stdout.
-s, --strict
Enable strict mode that checks for and reports duplicate fields in
entries.
COMMON OPTIONS
--help[=FMT] (default=auto)
Show this help in format FMT. The value FMT must be one of auto,
pager, groff or plain. With auto, the format is pager or plain
whenever the TERM env var is dumb or undefined.
--version
Show version information.
EXAMPLES
Deduplicate a single file:
$ bibdedup bibliography.bib -o clean.bib
Deduplicate multiple files using DOI:
$ bibdedup --keys doi file1.bib file2.bib -o merged.bib
Read from stdin:
$ cat input.bib | bibdedup -
Combine stdin with files:
$ cat extra.bib | bibdedup - existing.bib -o output.bib
Deduplicate using citekey with interactive conflict resolution:
$ bibdedup --keys citekey --interactive *.bib -o output.bib
Deduplicate with strict mode (reports duplicate fields):
$ bibdedup --strict --keys doi file1.bib file2.bib -o clean.bib
Deduplicate and output to stdout:
$ bibdedup --keys title,year file1.bib file2.bib
EXIT STATUS
bibdedup exits with:
0 on success.
123 on indiscriminate errors reported on standard error.
124 on command line parsing errors.
125 on unexpected internal errors (bugs).
BUGS
Report bugs to https://github.com/mseri/doi2bib/issues

Examples

doi2bib Examples

Read index entries from standard output and produce bibtex entries (one or more at a time):

$ doi2bib 10.1007/s10569-019-9946-9
$ doi2bib 1902.00436 arXiv:1609.01724 PMC2883744

Save bibtex entry to a file:

$ doi2bib doi:10.4171/JST/226 -o bibliography.bib

This will create the file if not present or append the bibliography to the existing file.

You can batch-process lists of entries by listing them line by line in a file and using the -i,--input option. For instance,

$ cat dois.txt
10.1007/s10569-019-9946-9
1902.00436
arXiv:1609.01724
PMC2883744
$ doi2bib -i dois.txt

bibfmt Examples

Format a bibtex file and print to stdout:

$ bibfmt bibliography.bib

Format a bibtex file and save to a new file:

$ bibfmt messy.bib -o clean.bib

Format bibtex content from stdin, using - as the filename:

$ echo"@article{key, title={My Title}, author={John Doe}}"| bibfmt -

Format with strict mode to check for duplicate fields (these can be removed with bibdedup):

$ bibfmt bibliography.bib --strict -q

You can use quiet mode to suppress normal output and only see warnings/errors:

$ bibfmt messy.bib --quiet

Force formatting even with parsing errors, by removing all the problematic entries (only do this after careful consideration):

$ bibfmt problematic.bib --force -o partial.bib

bibdedup Examples

Deduplicate entries from multiple files:

$ bibdedup file1.bib file2.bib -o merged.bib

Use custom keys for duplicate detection:

$ bibdedup --keys doi papers1.bib papers2.bib -o output.bib
$ bibdedup --keys title,year lib1.bib lib2.bib -o combined.bib

Deduplicate using citation keys:

$ bibdedup --keys citekey old.bib new.bib -o updated.bib

Interactive mode for conflict resolution:

$ bibdedup --interactive --keys title,author,year *.bib -o curated.bib

Enable strict mode to check for duplicate fields:

$ bibdedup --strict --keys doi papers.bib -o clean.bib

Installation

Each release comes with attached binaries for Windows, Mac, and Linux. You can simply unpack the binaries (doi2bib or bibfmt) and place them in a folder accessible by your terminal.

Building from Source

To build the package yourself, use opam:

$ opam install doi2bib # or bibfmt if you only need the pretty printer

This will install both doi2bib and bibfmt tools, since the latter is a dependency of doi2bib.

To run the tests, clone this repository and from the root of the project run:

$ opam install --deps-only .# first time only
$ dune runtest

Troubleshooting

If on macOS you get a Library not loaded: /usr/local/opt/gmp/lib/libgmp.10.dylib failure, you will need to install gmp:

• MacPorts users: port install gmp
• Homebrew users: brew install gmp

Editor Integration

Zed Configuration

Use the following to configure bibfmt as your bibtex formatter in Zed:

"languages": {
"BibTeX": {
"formatter": {
"external": {
"command": "/path/to/bibfmt",
"arguments": ["-"]
}
}
}
}

Replace /path/to/bibfmt with the actual path to your bibfmt binary.

Other Editors

Since bibfmt reads from stdin and writes to stdout by default, it can be easily integrated with other editors that support external formatters. The tool will preserve the content if parsing errors are encountered, making it safe to use in automated workflows.

API References

• DOI Content Negotiation
• arXiv API
• PubMed API

Development note

doi2bib was initially built by hand, with an initial working library before AI tooling was introduced. From that point, development was heavily assisted by Claude (via Copilot) accelerating the addition of features and tests, helping with the README and cli interface. The design, knowledge and direction remain the author's own; Claude was used as an accelerant, not an author.

I saw a similar disclaimer on kuva's repo (great project!) and decided to add it here as well, since it reflects how this project has evolved.