EndNote Field Code Converter

A ChatGPT Skill and standalone Python utility for converting Microsoft Word manuscripts with plain-text numeric citations into EndNote-ready .docx files.

The converter replaces ordinary bracketed citations such as [1], [1,2], and [1-3] with Microsoft Word field-code sequences using ADDIN EN.CITE. It also converts the numbered reference list into an ADDIN EN.REFLIST bibliography field. The output is designed to be opened in Microsoft Word with EndNote installed, so the references can be updated, formatted, and managed through EndNote.

What this project does

This project helps when a manuscript contains references only as plain text but needs to be converted into a Word document that behaves like an EndNote-linked manuscript.

Typical use cases include:

• converting journal manuscripts from plain-text Vancouver-style citations to EndNote-compatible fields;
• recovering an EndNote-like traveling library from a numbered reference list;
• preparing biomedical or scientific manuscripts for citation updating and bibliography formatting in EndNote;
• performing fast offline conversion without DOI lookup;
• optionally enriching parsed references with CrossRef DOI metadata.

Repository contents

.
├── LICENSE
├── README.md
├── RELEASE_NOTES.md
├── PUBLISH_TO_GITHUB.md
├── skill.zip
└── endnote-fieldcode-converter/
    ├── LICENSE
    ├── SKILL.md
    ├── agents/
    │   └── openai.yaml
    ├── references/
    │   ├── usage.md
    │   └── validation.md
    ├── requirements.txt
    └── scripts/
        └── endnote_converter.py

Installation

Use as a ChatGPT Skill

1. Download skill.zip from this repository.
2. Open the ChatGPT Skills page.
3. Upload skill.zip.
4. Use the Skill on a .docx manuscript that contains bracketed numeric citations and a numbered reference list.

Use as a standalone command-line script

Clone the repository and install the Python dependency:

git clone https://github.com/YOUR-USERNAME/endnote-fieldcode-converter.git
cd endnote-fieldcode-converter
python3 -m pip install -r endnote-fieldcode-converter/requirements.txt

Run the converter offline without CrossRef lookup:

python3 endnote-fieldcode-converter/scripts/endnote_converter.py \
  -i manuscript.docx \
  -o manuscript_EndNote.docx \
  --skip-crossref

Run with CrossRef metadata lookup:

python3 endnote-fieldcode-converter/scripts/endnote_converter.py \
  -i manuscript.docx \
  -o manuscript_EndNote.docx

Run a dry check without writing an output file:

python3 endnote-fieldcode-converter/scripts/endnote_converter.py \
  -i manuscript.docx \
  --dry-run \
  --skip-crossref

Supported citation patterns

The converter supports common bracketed numeric citation forms:

[1]
[1,2]
[1, 2]
[1-3]
[1,3-5]

By default, the script converts only citation numbers that are found in the parsed reference list. This reduces accidental conversion of bracketed numbers that are not citations, such as years or scale scores.

To force conversion of all numeric bracket patterns, use:

--convert-unknown-citations

Expected manuscript format

The manuscript should include a references section with a clear heading such as References, Reference, or Bibliography.

Example:

References
[1] Smith J, Doe A. Example article title. Example Journal. 2020;12(3):45-50.
[2] Kim H, Lee S. Another article. Journal Name. 2021;7:100-110.

In-text citations should appear before the reference section:

Previous studies reported similar findings [1]. The evidence is mixed [1,2] and has expanded recently [3-5].

Output

The output .docx preserves visible numeric citations while replacing their underlying Word XML with EndNote field-code structures:

w:fldChar begin
w:instrText ADDIN EN.CITE ...
w:fldChar separate
visible citation text, for example [1]
w:fldChar end

The reference list is replaced by an ADDIN EN.REFLIST field unless --keep-reference-text is used.

Command-line options

Option	Description
-i, --input	Input DOCX path
-o, --output	Output DOCX path. Defaults to input_EndNote.docx
--skip-crossref	Skip CrossRef DOI lookup. This is the fastest fully offline mode
--offline	Alias for --skip-crossref
--dry-run	Parse and validate without writing output
--convert-unknown-citations	Convert numeric brackets even if the citation number is not found in the reference list
--keep-reference-text	Keep the original plain-text reference entries after inserting the EndNote bibliography field

Validation checklist

After conversion, inspect the console output. A successful conversion should report:

• references found and parsed;
• converted citation count greater than zero;
• balanced Word field markers;
• output file written.

Treat these as issues requiring manual review:

• Field markers balanced: no
• Parsed references: 0
• Converted citations: 0 when the manuscript visibly contains [N] citations
• many Unparsed reference warnings
• many Skipped citation warnings for numbers that should exist in the reference list

Limitations

• The reference parser is heuristic and may not fully parse every journal style.
• EndNote must be installed in Microsoft Word to fully update, format, and manage the resulting fields.
• The script creates a traveling-library style XML payload from parsed reference text. It does not connect to a user's local EndNote library.
• CrossRef lookup may improve DOI and metadata fields, but it requires internet access and can fail for ambiguous titles.
• Complex citations inside Word text boxes, comments, footnotes, headers, or tracked changes may require additional OOXML handling.

Suggested GitHub repository description

ChatGPT Skill and Python script for converting plain-text numeric citations in DOCX manuscripts into EndNote field codes.

Suggested GitHub topics

endnote, docx, word, references, citations, manuscript, biomedical-writing, chatgpt-skill, ooxml, crossref

License

This project is released under the MIT License. See LICENSE.

Keywords / Search Terms

endnote zotero citation-manager reference-manager docx microsoft-word traveling-library addin-en-cite addin-en-reflist field-code crossref-api doi-lookup manuscript academic-writing biomedical vancouver-style citation-converter plain-text-citations zotero-to-endnote word-field-codes