ExoHub
The New Data Experience
Manage Data as Code. ExoHub standardizes how distributed datasets are versioned, synchronized, and governed across hybrid‑cloud and on‑premises systems using git‑based principles.
Harmonizing the Data Landscape
ExoHub brings architectural rigor to fragmented hybrid environments by treating data as code—normalizing access, versioning, and lineage across cloud object stores, sHPC, and SaaS. Built on top of git‑annex with unified credential handling, it eliminates unsynchronized copies and stale catalogs while ensuring reproducible, auditable datasets.
ExoHub Ecosystem
ExoGit — Versions massive datasets with git‑annex pointers for an auditable history while data lives across cloud and on‑premises backends.
exo CLI — Declarative manifests that standardize pull, sync, and publish workflows as code.
ExoGate — Secure on‑premises bridge so cloud workers can access protected sHPC/HPFS and S3 buckets without bespoke credentials.
ExoSpace — Geo‑localized cache that reduces egress and speeds compute by serving hot data locally.
ExoWorkers — Resilient orchestration (based on Temporal) with high‑throughput transfers (e.g., s5cmd) for reliable data publishing.
ExoSafe — Cloud-backed credential store that provisions SSH keys and git tokens in agentic, headless, or ephemeral environments.
ExoAtlas — Discovery and browsing layer for published datasets and artifacts, available both as the interactive `exo atlas` TUI and as a browser-based web interface.
Install exo CLI
One-line install for macOS/Linux
curl -fsSL https://github.com/Genentech/exohub/releases/latest/download/install.sh | bash