# PDF To Bibtex

> PDF To Bibtex is an Agent Skill for documents published in the JabRef/jabref repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Turns PDF papers into BibTeX entries with JabRef's jabkit CLI, pulling metadata from XMP, embedded BibTeX, GROBID, or text heuristics - builds a bibliography from a whole folder of papers.

- Skill name: `pdf-to-bibtex`
- Category: [Documents](https://www.openparable.com/skill/documents/1)
- Source: https://github.com/JabRef/jabref/blob/HEAD/skills/users/pdf-to-bibtex/SKILL.md
- Repository: JabRef/jabref (4,427 GitHub stars when collected)

## SKILL.md

# PDF to BibTeX

Extract bibliographic data from PDF files and produce BibTeX entries using `jabkit`, JabRef's command-line toolkit.

## Setup

`jabkit` runs via [JBang](https://www.jbang.dev/) — no JDK setup needed. Install JBang if missing:

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

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

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

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

One-off run without installation:

```bash
jbang --fresh jabkit@jabref --help
```

## Basic usage

```bash
jabkit convert --input paper.pdf --input-format pdfMerged --output paper.bib
```

Omit `--output` to print BibTeX to stdout. Add `--porcelain` before the subcommand for script-friendly output:

```bash
jabkit -p convert --input paper.pdf --input-format pdfMerged
```

## PDF importer formats

Pass one of these as `--input-format`:

| Format id | Strategy |
|---|---|
| `pdfMerged` | Merges results of the other importers into one best-effort entry — **use this by default** |
| `pdfXmp` | Reads XMP metadata stored in the PDF |
| `pdfEmbeddedBibFile` | Reads a BibTeX file embedded as PDF attachment |
| `pdfVerbatimBibtex` | Parses BibTeX printed verbatim on the first page |
| `pdfContent` | Heuristic extraction from the text of the first page |
| `pdfGrobid` | Sends the PDF to a [GROBID](https://github.com/kermitt2/grobid) service (requires GROBID to be enabled/reachable) |
| `pdfBibiliography` | Rule-based parsing of the bibliography section |

`--input-format "*"` auto-detects the format (also works for non-PDF inputs).

## Recommended workflow

1. Run with `pdfMerged` first.
2. Check the result for a `doi` field. If a DOI is present but fields look incomplete, fetch clean metadata instead:

   ```bash
   jabkit doi-to-bibtex 10.1145/3149935.3149942
   ```

3. Validate the resulting library:

   ```bash
   jabkit check paper.bib
   ```

## Batch conversion

```bash
for f in papers/*.pdf; do
  jabkit -p convert --input "$f" --input-format pdfMerged >> library.bib
done
```

Afterwards generate citation keys for all entries:

```bash
jabkit citationkeys generate library.bib --output library.bib
```

## Related

- Write metadata back into PDFs (reverse direction): `jabkit pdf update --citation-key <key> --input library.bib --input-format bibtex`
- Full CLI reference: the `jabkit` skill in this repository.

---

Source: https://www.openparable.com/skill/documents/pdf-to-bibtex (OpenParable). Cite the canonical URL above.