ePrint Backups
This site contains backups of the Cryptology ePrint Archive. A snapshot of ePrint is taken every month.
Download
October 2026:
September 2026:
Backup Structure
Each full snapshot only contains the latest papers as of the time the snapshot was taken. No paper has more than one version in a single snapshot.
The backup has the following structure:
cutoff.txt — The cutoff time (ISO 8601) of the backup contents.
metadata.jsonl — A JSONL file containing the metadata of every file in the backup, sorted by year then paper number, in ascending order (ie starting at 1996/001 and ending this year, and where 2026/999 preceeds 2026/1000). No paper ID appears twice. Each line is a JSON object with the following fields:
id — Paper ID, of the form <year>/<num>, where <num> is a decimal number, padded with up to 2 zeros.
oai_identifier — Paper identifier in the OAI metadata exposed by the site (eg "oai:eprint.iacr.org:2026/2237").
datestamp — The last time (ISO 8601) the OAI record was updated.
status — "active" (meaning the paper is not withdrawn) or "deleted" (meaning the paper has been withdrawn). If "deleted", all the following fields are null.
url — The URL to the paper on ePrint.
title — The title of the paper.
creators — A list of authors.
subjects — A list of subjects within cryptography that the paper pertains to.
abstract — The abstract of the paper. This is a string, usually but not always containing valid LaTeX. Sometimes it contains HTML.
date_submitted — The time (ISO 8601) the paper was submitted to ePrint.
date_last_revised — The time (ISO 8601) the paper was last edited. This can include edits to title, abstract, author list, or the document file itself.
rights — A URL linking to the document's copyright.
<year>/ — Directories named by year. Each directory has files of the following form:
<num>_<timestamp>{.pdf,.ps,.ps.gz} — Papers, named by the <num> component of their paper ID, followed by an underscore, followed by the Unix epoch time in seconds that the document was uploaded (or an estimate thereof), followed by the file extension. Examples: 016_850176000.ps, 1312_1783485936.pdf, or 003_1010503460.ps.gz.
As with any archive, some creative choices are necessary in the face of missing or incomplete data. Read Methodology to see how we handle this.
License
All papers in the backup are licensed with one of:
CC BY 4.0,
CC BY-SA 4.0,
CC BY-NC 4.0,
CC BY-NC-SA 4.0,
CC BY-NC-ND 4.0, or
CC0 Public Domain.
Each paper's license is indicated in its entry in metadata.jsonl described above.
All original work on this site and in the backups themeselves is freely given under CC0.
Contact
telluride11 AT proton.me