Use the RHEL command-line assistant offline with this new developer preview

Use the RHEL command-line assistant offline with this new developer preview Architecture overview Prerequisites and requirements Hardware requirements Getting started Configure the GPU Configure the command-line assistant client Usage First query response delay CPU-only systems Intended for individual system use cases Conclusion Get started with AI Inference About the authors Brian Smith Máirín Duffy Sally O'Malley More like this Blog post Blog post Original podcast Original podcast Keep exploring Browse by channel Automation Artificial intelligence Open hybrid cloud Security Edge computing Infrastructure Applications Virtualization Share An offline version of the Red Hat Enterprise Linux (RHEL) command-line assistant powered by RHEL Lightspeed is now available as a developer preview to existing Red Hat Satellite subscribers. This delivers the power of the RHEL command-line assistant in a self-contained format that runs locally on a workstation or an individual RHEL system. This enables the assistant to function in a completely disconnected, offline, or air-gapped environment because it doesn't require any external network connectivity to operate. The command-line assistant provides AI-powered guidance and suggestions to help you with a wide range of RHEL tasks. We’ve extended its knowledge about RHEL installation and upgrades topics, enabling you to get assistance in these areas even with limited or no connectivity. The offline version of command-line assistant is delivered as a set of container images that can be run with Podman. The containers used are: installer container : Pulls the other required containers, installs the rhel-cla command, and optionally creates a systemd service rlsapi container : Provides the endpoint that the command-line assistant client communicates with rag-database container : The retrieval-augmented generation (RAG) database used to supplement the LLM’s knowledge with additional data such as the RHEL documentation ramalama container : Provides LLM inference Your Red Hat Satellite subscription provides you with tools to locally manage your RHEL environments. These tools include the Satellite product itself, as well as the recently introduced Red Hat Offline Knowledge Portal , which provides an offline version of Red Hat's exclusive knowledge content. We've now expanded the Satellite subscription to also provide the offline version of the RHEL command-line assistant. The offline version of the command-line assistant is delivered as a set of containers, and can run on a RHEL system, or a Mac or Windows workstation. If you're using a Mac system, Podman Desktop is required , and it's recommended to use a system equipped with an NVIDIA, AMD, or Mac M-series GPU. CPU-only systems (RHEL 9.6+ and 10+ / Fedora 42 / Windows 11): RAM : 8 GB CPU cores : 2 GPU-capable systems (RHEL 9.6+ and 10+ / Fedora 42 / Windows 11): RAM : 4 GB GPU : At least 4 GB of VRAM Apple systems (macOS 15.