git-annex

Overview

git-annex manages large files for people who want to keep file contents out of Git while tracking filenames and metadata in repositories. It suits command-line-oriented Git users coordinating files across computers, devices and storage services. Its distributed design is flexible, but its own guidance is clear: it is not a backup system by itself.

Key features

Git holds filenames and metadata; annexed contents live in a distributed store across repository clones. That separation lets users sync, archive and move files without making Git carry the file contents. Transfers can resume after interruption, a practical safeguard for large or unreliable transfers.

The assistant can create a synchronized folder across computers, Android devices, removable drives, NAS appliances and cloud services. Built-in special remotes include S3, Glacier, rclone, rsync, web, WebDAV and Android adb, giving technically inclined users a broad choice of storage destinations.

Encrypted special remotes can encrypt contents and use HMAC hashing for filenames. Encryption is mostly not used by default, so users handling sensitive data need to configure an encrypted remote rather than assume stored files are protected.

File size is limited by the filesystem. Repositories with hundreds of thousands of files are supported, but larger ones may slow down because of Git limitations. The project says git-annex is not a backup system; it can instead form part of an archival workflow or deliver files to a separate backup system.

Pricing

git-annex is free and open source. The git-annex plan costs nothing and manages large files without storing their contents in Git. There is no paid tier or free-trial period to weigh; users provide or arrange the storage destinations themselves.

Platforms

Official installation instructions cover Linux distributions, macOS, Android and Windows, with package and other installation methods. The product is also categorized for self-hosted and web use. Its assistant's synchronized folders can span computers, Android devices, removable drives, NAS appliances and cloud services.

Who it's for

Choose git-annex if you already work with Git, are comfortable with the command line, and need to coordinate or archive large files across distributed storage. It is a poor fit as a standalone backup solution, or for someone who wants file protection without configuring storage and encryption choices.

It supports file-level snapshots, dataset branching and point-in-time rollback, with both storage backend types and a self-hosted deployment model. The project also notes that DataLad uses git-annex to provide access to scientific data from various sources.

Pros and cons

• Pros: Keeps large contents outside Git while preserving filenames and metadata, which avoids putting those contents directly in repository history.
• Pros: Supports many storage destinations and can resume interrupted transfers, useful for distributed workflows and lengthy transfers.
• Pros: Free, open source and installable across major desktop platforms and Android.
• Cons: Not a backup system on its own, so users must build a separate backup process where needed.
• Cons: Encryption is mostly not enabled by default, requiring deliberate configuration for sensitive contents.
• Cons: Very large repositories may slow down because of Git limitations, even though hundreds of thousands of files are supported.

Alternatives

For a broader category comparison, see Data Version Control Tools.

• Weights & Biases is a freemium alternative with a free plan and a free trial; its Free plan includes 5 model seats, 5 GB/month storage and 1 GB/month of Weave data ingestion.
• Apache Paimon is a free, open-source data lake platform; choose it when you need to select connectors and storage plugins for your environment.
• Dolt is a free, open-source option with a roughly 100 MB program; consider it as another Git-oriented data tool.
• ClearML offers a freemium model and says its self-hosted version is 100% open source on GitHub; consider it when a self-hosted open-source option is the priority.
• Neon has a free tier with 10 projects, 50 CU-hours per month per project, 0.5 GB storage per project, 10 branches per project and 5 GB monthly egress; consider it when those database project limits suit your needs.
• Nile offers a free local plan that runs entirely on your machine without a cloud account.
• DataLad is free and open source, and is a related option for scientific data workflows; the DataLad project uses git-annex.
• Hugging Face Hub has a free plan with 100 GB private storage and best-effort public storage; consider it when hosted model and dataset storage is a better fit.

Verdict

git-annex is a strong choice for command-line Git users who need distributed, resumable handling of large files without putting contents in Git. Its range of remotes and free, open-source model are compelling; look elsewhere or pair it with a separate system if you need a backup solution or encryption enabled by default.