How to create an anonymous repository for double-blind peer review?

[asmita7798]

Select Topic Area

General

Body

Hello everyone,
I am preparing a research paper submission that requires providing reviewers with access to an anonymous code repository (as part of double-blind peer review). The guidelines specify that the repo link should not contain any identifiable information about the authors.

I’m a bit unsure how best to set this up on GitHub:

1. Is there a way to create a repository that does not expose my username or personal details?
2. Are there recommended practices researchers typically follow in such cases?
3. If GitHub isn’t suitable for this purpose, I’d also appreciate suggestions for alternative platforms commonly accepted by conferences/journals.

Thanks in advance for any guidance!

[softwarecreations]

Creating an Anonymous GitHub Repo for Double-Blind Review

GitHub doesn't natively support fully anonymous repositories (your username is always visible in the repo URL and metadata), but you can minimize identifiability with these steps:

1. Use a throwaway account: Create a new GitHub account with no personal details (e.g., generic email like anon-reviewer@example.com). Avoid linking it to your real profile. Push your code to a new repo under this account, naming it something neutral like anonymous-project-2025.

2. Sanitize the repo:

   • Remove any author names, emails, or affiliations from README.md, commit messages, code comments, and metadata (e.g., via git filter-branch to rewrite history).
   • Don't include links to your personal sites or other repos.

3. Share access: Provide the repo URL in your submission. For extra caution, make it public (reviewers expect this) or invite reviewers via temporary emails if private.

Recommended practices: This is standard in academia. Many use throwaway accounts for NeurIPS/ICML/etc. submissions. Always check conference guidelines (e.g., some allow OpenReview for anon links).

Alternatives:

• GitLab or Bitbucket: Similar to GitHub; create anon accounts.
• Zenodo or Figshare: For read-only archives with DOIs, fully anonymous uploads.
• Anonymous Git hosting: Services like anonscm or self-hosting on a VPS, but less common.

This keeps your identity hidden while enabling review. If your conference specifies tools, follow those first!

Was this helpful?
If so, please mark this as the answer. I would appreciate it.

[Ahrakshith]

Hey! This is a pretty common situation for research submissions that use double-blind review. Since GitHub accounts are tied to your username and profile, there’s no built-in way to completely hide your identity on your main account. Check out anonymous.4open.science
,it creates an anonymous mirror of your GitHub. [repo.https://anonymous.4open.science/ ]

[kslote1]

Just a personal note from trying to use this: I do not recommend using this service at this time.

There are bugs in the software that caused google chrome to glitch out. The bugs made me uncertain about trusting it with my scientific software or giving this app access to my github orgs.

It's most likely better practice to use OSF or Zenodo.

https://osf.io/
https://zenodo.org/

[whatsupsumit]

Hello,

For double-blind peer review, the goal is to provide access to your code without revealing your identity. Here are some best practices and options:

Using GitHub Anonymously

Create a new account that doesn’t include your name, email, or any other identifiers. Use a neutral username like anonymous-researcher123.

Set the repository to private and then generate a shareable link using GitHub’s “Invite via link” feature. Only reviewers with the link can access it.

Avoid committing any files with metadata that could reveal your identity (e.g., author info in scripts, notebooks, or commit history). Consider squashing commits or using git commit --author="Anonymous anon@anon.com".

Other Recommended Platforms

Zenodo (https://zenodo.org
): Free, DOI-enabled repository. You can upload your code anonymously, and it’s widely accepted in conferences/journals.

OSF (Open Science Framework) (https://osf.io
): Allows anonymous project creation and controlled access for reviewers.

Figshare (https://figshare.com
): Another repository where you can share materials anonymously.

Additional Tips

Remove identifying information from file names, code comments, and documentation.

Include a README explaining how to run the code without any personal references.

Use DOIs or temporary links instead of your personal GitHub URL if the conference explicitly forbids identifiable URLs.

Summary:
For GitHub, the key is creating a neutral account and private repo. For a more formal or DOI-backed solution, Zenodo or OSF are excellent choices and often preferred by conferences.

[MuhammadAlZafarKhan]

This is a super common issue, and many people wonder about how to do this, but never ask.

Creating anonymous repos are very common when sunbmitting to major AI conferences like NeurIPS.

• The most straightforward way to create an anonymous repository for double-blind peer review is to use a dedicated anonymization service like Anonymous GitHub (anonymous.4open.science), or to submit your code as an anonymized ZIP file through the conference's supplementary material system.

• Follow these steps for Anonymous GitHub (anonymous.4open.science):

1. Authenticate with your personal GitHub account.
2. Provide the URL of your repository.
3. You will be prompted to list any terms that could reveal your identity, institution, or personal details. This list typically includes: Your name, usernames, and email addresses, your institution or company name(s), and any specific internal project names that could trace back to you.
4. The service creates a mirror of your repository on their platform, automatically replacing the specified terms (and other identifying information like the repo owner/organization name) with placeholders like "XXXX."
5. Include the resulting anonymous URL (e.g., $\text{anonymous.4open.science/r/[unique-id]/}$) in your paper and/or supplementary files as instructed by the conference.

[antipodite]

slop response