refren: scientific manuscript PDF renamer
Scenario: when you download a scientific paper from the Internet, the PDF file is usually named someting cryptic like 2506.19540v1.pdf
refren renames it to FirstAuthor_SecondAuthor_JournalAbbrev_Year.pdf using Claude AI to extract bibliographic metadata from the PDF
Conveniently, it can also download the supplemental data for the manuscript automatically
Usage
refren <pdf_file> [--remove] [--supplement]
--remove deletes the original PDF after creating the renamed copy.
--supplement also searches the web for the paper's supplementary material and downloads it. If a supplement is found and successfully downloaded, refren creates a FirstAuthor_SecondAuthor_JournalAbbrev_Year/ folder and puts both the renamed PDF and the downloaded supplement inside it. If no supplement can be found or downloaded (e.g. the publisher blocks automated downloads), refren falls back to the normal behavior: just rename the PDF in place, no folder created.
--supplement reliability by source
Many publishers block automated downloads outright, so this is best-effort, not guaranteed. As last verified:
| Source | Status | Notes |
|---|---|---|
| arXiv | ✅ Works | Checked directly via arXiv's own ancillary-files listing, not a web search |
| Nature-branded titles (nature.com) | ✅ Works | Nature, Nature Medicine, Nature Biotechnology, Nature Catalysis, Nature Energy, Signal Transduction and Targeted Therapy, etc. — anything with a 10.1038 DOI, checked directly via nature.com. (Springer Nature also publishes non-Nature-branded journals on a different platform — see below.) |
| Springer/BioMed Central titles (link.springer.com, biomedcentral.com) | ❓ Unknown | Untested as a deterministic check — a different site than nature.com, despite shared corporate ownership. One example (J Cheminform) hit a separate issue: its article page is a JS-rendered SPA with no text in the raw HTML |
| PubMed Central (PMC) | ❌ Doesn't work | Article pages are readable, but file downloads are blocked by NCBI's bot-detection challenge |
| Elsevier / ScienceDirect | ❌ Doesn't work | Site returns 403 to automated requests outright |
| The Lancet | ❌ Doesn't work | Site blocks automated access; recent trials checked had no PMC copy to fall back to either |
| Annals of Oncology | ❌ Doesn't work | Same ScienceDirect/Elsevier block; recent papers checked had no PMC copy either |
| NEJM | ❌ Doesn't work | Site returns a Cloudflare bot-challenge page instead of the file |
| Other publishers | ❓ Unknown | Untested |
Examples
$ refren 1758-2946-6-10.pdf
(calling Claude API...)
First author last name : Krstajic
Second author last name: Buturovic
Journal : Journal of Cheminformatics -> J Cheminform
Year : 2014
1758-2946-6-10.pdf -> Krstajic_Buturovic_JCheminform_2014.pdf
Copied to: Krstajic_Buturovic_JCheminform_2014.pdf
$ refren 2608.14705v1.pdf --supplement
(calling Claude API...)
Article type : research
First author last name : Buturović
Second author last name:
Journal : arXiv -> arXiv
Year : 2026
(checking arXiv for ancillary files...)
2608.14705v1.pdf -> Buturović__arXiv_2026/Buturović__arXiv_2026.pdf
Supplement saved to: Buturović__arXiv_2026/
arXiv-2608.14705v1.tar.gz
Install
$ pipx install refren # Linux, Mac
Requires an ANTHROPIC_API_KEY environment variable. If you don't have the API key, please read SETUP.md for instructions.
Development
cd ~/github/refren
rm -f dist/*
uv run python -m build
uv run twine upload dist/*
1 maintainer
Metadata
1 maintainer
Release files for refren 0.18.0
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| refren-0.18.0.tar.gz | 13.5 MB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| refren-0.18.0-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 13.5 MB
Release files / refren-0.18.0.tar.gz
| Download URL | refren-0.18.0.tar.gz |
|---|---|
| Size | 13.5 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
a69fa936528b2480a3b93a71e09e6f23d8df0dfdcc76ec6588b946fc55f9e885
|
|
BLAKE2b-256 checksum How to use checksums |
848a377e28e046118ebe18743d1e80b0ceef35cde4d7cdd13f03089931e70f85
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.4
|
Release files / refren-0.18.0-py3-none-any.whl
| Download URL | refren-0.18.0-py3-none-any.whl |
|---|---|
| Size | 11.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
353237b63c27ee0f1a84358f6f38575acaca0937ad0186b1127b6ce9be0cc7de
|
|
BLAKE2b-256 checksum How to use checksums |
c5bebd8101072de7ed982d6fb92cdcc87e15cb2149d35d17b052077c81ff75a5
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
uv/0.8.4
|
Security for refren 0.18.0
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
