Unleash the Compute.

Atra — the Swiss-grade supercomputing stack for AI & Research. Bare-metal performance. Cloud agility. No hypervisor tax. Get your workloads off the cloud and onto the rock.

Deploy Cluster Read Documentation
user@scops-lab:~$ atra list --clusters
> Cluster: Omega-1 | Status: READY | Purpose: AI Research | Nodes: 450 | GPUs: 1800 | Storage: 1.8 PB | Network: 200 Gbps
user@scops-lab:~$ atra list services --cluster=Omega-1
> List of services for cluster Omega-1: SSH, Storage, Container, Slurm, Security
user@scops-lab:~$ atra list resources --cluster=Omega-1
> nid000[0-450] compute nodes | Type: GH200 | 4 GPUs each.
_
atra logo
OPEN-SOURCE FRAMEWORK

Atra

Atra bridges bare-metal HPC infrastructure with cloud-style service composition. Scientific communities deploy tailored, high-performance computing environments on shared supercomputers — without virtualization.

Bare-metal

Physical hardware. Zero virtualization overhead.

Composition

Services layered and assembled like building blocks.

Immutability

Deployed environments are locked and reproducible.

Scientific Power

Built for HPC researchers, weather forecasting, particle physics, ML.