Anonymous GitHub

Share your code and data for double-anonymous peer review — without giving your identity away.

Public instance · How it works · Self-hosting

Why

Double-anonymous review asks you to anonymize the artifact behind your paper — the code and data — exactly like the paper itself. Doing that by hand (owner, organization, repository name, logins, and every mention buried in the files) is tedious and easy to get wrong. Anonymous GitHub does it for you and serves the result behind a shareable link.

A free public instance is available at anonymous.4open.science.

What it anonymizes

• The GitHub owner, organization, and repository name
• File and directory names
• File contents of every extension (Markdown, text, source code, …), replacing a configurable list of terms

Usage

Public instance

Just open anonymous.4open.science, paste your repository URL, add the terms to hide, and share the generated link.

CLI

Anonymize a repository locally and produce an anonymized zip:

npm install -g @tdurieux/anonymous_github
anonymous_github

Self-hosting

Local installation requires Node.js 22.15 or newer in the 22.x series, 24.x, or 26 or newer. Run npm ci and npm run build to install and build. npm run dev uses Node's watch mode; npm run build:ui rebuilds browser assets.

Run your own instance with Docker

1. Clone and install

git clone https://github.com/tdurieux/anonymous_github/
cd anonymous_github
npm i

2. Configure GitHub access — create a .env file:

GITHUB_TOKEN=<GITHUB_TOKEN>
CLIENT_ID=<CLIENT_ID>
CLIENT_SECRET=<CLIENT_SECRET>
CREDENTIAL_KEYS='{"2026-09":"<base64-encoded 32-byte random key>"}'
CREDENTIAL_ACTIVE_KEY_ID=2026-09
CREDENTIAL_LEGACY_READS=false
PORT=5000
DB_USERNAME=
DB_PASSWORD=
AUTH_CALLBACK=http://localhost:5000/github/auth

• GITHUB_TOKEN — create one at https://github.com/settings/tokens/new with the repo scope.
• CREDENTIAL_KEYS / CREDENTIAL_ACTIVE_KEY_ID — generate a key with openssl rand -base64 32. Existing installations must follow the credential migration guide before starting this release.
• CLIENT_ID / CLIENT_SECRET — from an OAuth App at https://github.com/settings/applications/new.
• The App's callback must be https://<host>/github/auth (matching AUTH_CALLBACK).

To enable read-only access to selected private repositories alongside OAuth, follow the GitHub App setup guide.

3. Start the server

docker-compose up -d

4. Open http://localhost:5000. The port can be changed in docker-compose.yml; putting Anonymous GitHub behind nginx is recommended for HTTPS.

For an optional remote, hidden, delayed MongoDB replica and backup source, see the MongoDB replication guide.

Owner email alerts

Set RESEND_API_KEY and EMAIL_FROM on the web server and download workers to send access alerts through Resend. Use a sender address on a domain verified in Resend. APP_HOSTNAME sets the host used for the connections link in emails.

Registration saves a verified GitHub email address, preferring the primary address. OAuth requests user:email; GitHub App sign-in uses the read-only email permission described in the setup guide. Later sign-ins also fill missing addresses. Saved addresses and custom alert addresses are preserved. GitHub privacy placeholder addresses are not imported.

Signed-in owners without an email address see an email form in a popup. “Remind me later” dismisses it for the current login session; “Never” saves a permanent preference on the account. Existing addresses are reused, and owners can add or change their address on the connections page at any time. Alerts cover invalid GitHub credentials and repository access errors encountered while reading or refreshing content. No periodic access scan is performed. Emails contain no repository names, anonymous URLs, or credentials.

An atomic recipient claim allows one email attempt across repositories, pull requests, gists and accounts sharing the same address. Sign-in does not reopen alerts. An explicit reconnect or resource change prepares recovery, but the recipient stays blocked until a fresh GitHub read confirms access to the resource that triggered the alert. A fresh read during an in-flight send is retained and applied when that send completes. A newer reset or source change invalidates that recovery evidence. Cached reads and successful reads of other resources cannot reopen it. Removed or expired resources and disabled owners no longer block future alerts after their send has completed.

Dispatched recipient guards have no time expiry, so there are no daily reminders. Failed deliveries remain blocked until confirmed recovery. An interrupted send whose outcome is unknown stays blocked to avoid duplicate delivery. Unsent claims can be cancelled by confirmed recovery before dispatch. Bookkeeping failures release claims when no provider request has started. Pending claims interrupted before dispatch can be reclaimed after a 60-second lease during startup or a new notification attempt. Claim identity and state checks prevent a paused sender from dispatching after another instance reclaims its claim. Cancellation is saved before cleanup; matching legacy fields are cleared before the guard is released. Interrupted cleanup resumes on startup or a later notification. Legacy fields carry an attempt ID so unrelated deliveries and newer generations are preserved. A resource revision prevents delayed preparation from recreating a reclaimed claim. All asynchronous preparation stays pending. A sending reservation is reclaimable until dispatch commits atomically on the owner document, against its reset revision and lease, immediately before contacting the provider. Cancellation competes with that same commitment, so a paused sender cannot dispatch afterward. Owner dispatch records distinguish known-unsent reservations from interrupted committed sends. Completed records are removed; expired uncommitted records are pruned on later attempts. Persistent reset revisions also fence attempts created while a reset is in progress, even when no recipient guard existed at its start. Recovery recorded during an active reset is retried when reset finishes or its lease expires, with the original source and revision checks. Provider idempotency keys and redacted attempt records support auditing without storing recipient addresses or repository names in the audit log. Attempt records are saved before dispatch, with completion recorded separately. Existing resource claims are migrated to recipient guards before new application instances serve requests using an indexed scan with resumable checkpoints. Other instances wait while its lease is healthy, and take over if the lease expires. Every existing resource claim is retained behind the recipient guard. Recovering one resource atomically marks its claim recovered; other claims keep the recipient blocked until they recover or are retired. A newer owner reset makes its matching claims require another fresh read. Completed provider results are recorded before guard updates, with bounded retries and reconciliation on startup or a later notification. Retirement also checks the latest owner reset revision.

Use scripts/deploy.sh for the first upgrade to recipient alerts. The new image has the science.4open.owner-alert-protocol=recipient-v1 label. When running application containers lack that label, the script builds the replacement, stops and verifies all old API and streamer containers, then prepares a full claim reconciliation and starts the new streamers and API. Stopped legacy containers from a rollback also trigger reconciliation, so claims written after an earlier migration are included. This path also runs when retrying with an already-built image and FORCE=0. Once writers are stopped, preparation releases all known unsent pending claims and clears their matching legacy fields before backfill. This upgrade pauses traffic while claims migrate and allows up to 15 minutes for startup, configurable with MIGRATION_HEALTH_TIMEOUT. Subsequent deployments keep the usual rolling streamer replacement. Manual upgrades must stop all legacy application writers, run docker compose run --rm --no-deps --entrypoint node anonymous_github ./build/scripts/reconcile-legacy-owner-alerts.js using the new image, and then start the new application instances. Run this preparation after legacy writers have operated, even if their containers were already removed. Ordinary restarts preserve confirmed recoveries. Leaving either Resend setting empty disables delivery and the email prompt.

Scope of anonymization

In double-anonymous review, the boundary of anonymization is the paper plus its online appendix — and only that. Googling part of the paper or appendix to reveal authorship is considered a deliberate attempt to break anonymity (explanation).

How it works

Anonymous GitHub either downloads the full repository and anonymizes each file, or proxies requests to GitHub on the fly. In both cases the original and anonymized versions are cached on the server, so even large repositories stay responsive.

Related tools

• gitmask — contribute anonymously to a GitHub repository.
• blind-reviews — browser add-on that hides identifying information when reviewing a pull request.

See also

• Open science and double-anonymous peer review
• ACM policy on double-blind reviewing

License

GPL-3.0 © Thomas Durieux

Frontend development

The UI uses Vue 3 and Vue Router with the existing Express API. Page setup functions live in public/script/app.js and admin.js; the Vue templates are in public/partials/. Gulp compiles the templates and bundles the app with esbuild, then updates the asset manifest used by Express.

Run npm run build:ui after changing a template or frontend script. Use npm run dev:ui to serve the built UI locally with API requests proxied to the configured upstream. npm run test:ui rebuilds the assets and runs the frontend regression and DOM interaction tests. npm run build also builds the UI for production.

The initial bundle contains the Vue app. Markdown extensions load on content routes; PDF.js, Ace, and notebook support load when their viewers mount. Org support loads in the repository explorer, and Mermaid loads only when a diagram is encountered. These libraries use hashed URLs and load once per tab.

The Docker build compiles the same bundles and copies them, the asset manifest, and document-worker assets into the runtime image. The Compose app serves these assets directly; no frontend development server is needed.