paperblind serves PDFs, supplementary material, and demos for academic papers while they are under double-blind peer review. Each paper lives at its own opaque URL, protected by a per-paper password shared with reviewers.
Have a review link? Open the URL exactly as given and enter the password it came with. There is no public index.
Each paper is served from an unguessable slug. The landing page lists nothing; only invited reviewers hold the link.
Access is enforced per paper with scoped cookies. Wrong passwords and unknown slugs look identical from outside.
Outbound links and metadata are rewritten so proxied content cannot leak author homepages, affiliations, or analytics.
No analytics, no third-party scripts, no public enumeration of hosted papers. HTTPS with HSTS; minimal headers.
Each blinded slug maps to a paper page, which can be password protected, on an origin server. paperblind fetches from that origin and serves the response through the blinded URL.
Papers with Code's practical checklist for releasing reproducible artifacts alongside a submission.
Widely-referenced guidance on what does and does not count as identifying information in an anonymous submission.
Double-blind process details, anonymity expectations, and supplementary material handling across venues.
Community service for sharing GitHub repositories with reviewers while hiding author identity.
Create a standalone mirror under a neutral account without carrying over fork graphs or contributor history.
Push a detached anonymous remote that shares no commit history with your identifiable working repository.
Avoid hostnames, usernames, or project names that can be traced back to authors or their institutions.
Strip authorship from PDF properties, Office docs, and image EXIF before uploading supplementary material.
Remove links to personal homepages, lab pages, and profile URLs — even one can deanonymize a submission.
Third-party scripts and embeds can leak account IDs, organization identifiers, or referrer data to reviewers.