↓Skip to main content
  1. Infrastructure/

GIN (G-Node Infrastructure)

·3 mins·
Ai-Partial Infrastructure Git-Annex CON-Contrib Git-Annex Gogs Self-Hosted Forge Neuroscience Research-Data
Language: Go
License: MIT
Maturity: stable
Verified: 2026-02
Examples: gin.g-node.org (public instance)
Table of Contents

Overview #

GIN (G-Node Infrastructure) is a research data management platform built on a fork of Gogs with added git-annex support. It is operated by the German Neuroinformatics Node (G-Node) at Ludwig Maximilian University Munich and provides free data hosting for neuroscience researchers.

GIN and Forgejo-Aneksajo address the same fundamental problem – standard git forges cannot handle git-annex content – but arrived at it from different directions and lineages:

  • Gogs (2014) was the original lightweight Go git forge
  • Gitea (2016) forked from Gogs with a more open governance model
  • Forgejo (2022) forked from Gitea after governance concerns
  • GIN forked Gogs directly, adding git-annex support
  • Forgejo-Aneksajo added git-annex support to Forgejo

GIN was funded by the same NSF+BMBF CRCNS (Collaborative Research in Computational Neuroscience) program that funded DataLad, reflecting the shared recognition that neuroscience needs proper infrastructure for large research data.

Key Features #

  • git-annex integration – large file versioning via git-annex, transparent to users through the web interface and GIN CLI
  • DOI registration – publish datasets and receive a DOI for citation
  • Web interface – browse repositories, file trees, and commit history
  • Private and public repos – access control for sensitive data with the option to publish when ready
  • GIN CLI – dedicated command-line client (gin upload, gin download, gin get) that wraps git and git-annex operations into simpler commands

GIN vs Forgejo-Aneksajo #

GINForgejo-Aneksajo
Base forgeGogsForgejo
git-annex supportYesYes
Active upstreamGogs is less actively developedForgejo is very actively developed
Known instancesgin.g-node.org (single public instance)hub.datalad.org, hub.psychoinformatics.de, hub.edu.datalad.org
Self-hostingPossible but uncommonDesigned for self-hosting
DataLad integrationPartial (git-annex works, datalad push works)Native (datalad create-sibling-gogs)
CommunityPopular among EU neuroscience academicsGrowing DataLad/research data community
FundingBMBF (German federal)Community-driven

git-annex / DataLad Integration #

Integration level: git-annex.

GIN supports the git-annex protocol, so git annex copy --to and git annex get --from work against GIN repositories. DataLad datasets can use GIN as a sibling:

# GIN works as a standard git remote with annex support
datalad create-sibling-gogs --name gin \
    --api https://gin.g-node.org/api/v1 \
    --credential gin-token

datalad push --to gin

The GIN CLI provides a simplified interface that hides git-annex complexity:

gin login
gin create my-dataset
gin upload .
gin download my-dataset

AI Readiness #

Level: ai-partial.

Repository metadata and file listings are accessible via the Gogs API. Actual data content depends on what researchers upload – structured formats (NIfTI, BIDS) are AI-processable, while raw binary data may need domain-specific tooling.

See Also #

  • Forgejo-Aneksajo – the more actively maintained alternative with similar goals
  • DataLad Hub – a Forgejo-Aneksajo deployment for DataLad datasets