This page outlines the recommended install paths in 2026.
If you only want to run remote workflows, you can often skip most of this and start at:
Option A (recommended for most users): container image
If your environment has Singularity or Apptainer, the container image is the lowest-friction way to get a working CLI.
wget "https://b2drop.bsc.es/index.php/s/ay9mqyD8CTE3GgA/download" -O ~/rbbt.singularity.img
# Singularity
alias rbbt='singularity exec -e ~/rbbt.singularity.img rbbt'
# Apptainer (alternative)
# alias rbbt='apptainer exec -e ~/rbbt.singularity.img rbbt'
rbbt --help
When using the container, Rbbt will still create state under your home directory (cache, resources, etc.).
Option B: developer install (Ruby gems)
If you plan to author workflows, contribute to the framework, or need a fully local setup, you’ll likely want a Ruby-based install.
High-level steps:
- Install a Ruby version compatible with your Rbbt/Scout deployment.
- Install the required gems.
- Configure paths (resources, cache, workflows).
This repo currently contains the historical installation tutorial:
(As the docs are modernized, this page will be expanded with more explicit “developer” vs “user” setup guidance.)
Option C: Docker / VM provisioning
Rbbt has tooling to help with provisioning images/VMs in some deployments.
See:
Common setup concepts (regardless of install)
Bootstrapping
The first time you run a workflow task, it may need to download/build indices or install third-party tools. This can take time.
Caching / persistence
Rbbt/Scout heavily relies on caching so results are reproducible and incremental. You will typically see job directories (“steps”) created under your working directory or a configured cache dir.