Why Ariax
Choose the workflow that fits the design, move from a prepared target to ranked candidates, and pay only for the GPU time your campaign uses.
01Workflow choice
Ariax makes state-of-the-art protein-design methods accessible through guided interfaces while keeping the underlying scientific controls visible.
Start with the design objective—not infrastructure. Prepare the target, configure the experiment, follow each stage, and keep the resulting structures and scores connected to the project that produced them.
Compare workflow capabilities02Working model
The meaningful difference is not a checklist of infrastructure features. It is how directly you can move from scientific intent to a traceable protein-design campaign.
| Design workflow | Ariax Bio | DIY cloud / HPC | Google Colab | Other software platforms |
|---|---|---|---|---|
| Workflow choiceChoose the scientific method that fits the design objective. | BindCraft, BoltzGen, PXDesign, and the ESMFold2 pipeline | Assembled and maintained by your team | Notebook-dependent | Platform-dependent |
| Project setupTranslate a target and design intent into a reproducible campaign. | Guided setup built around each workflow | CLI, configuration files, and environment setup | Notebook cells and session state | Platform-specific |
| Starting a runMove from prepared inputs to active design work. | Launch without a shared queue or quota request | Depends on your cluster or cloud account | Subject to session and runtime limits | Platform-dependent |
| ParallelizationCompress campaign time by running design work across multiple GPUs. | Turbo Mode: 2×, 4×, or 8× parallel execution across every engine | Possible, but requires multi-GPU orchestration and additional setup | Not available | Not standard; availability and controls vary |
| Campaign visibilityStay close to every stage of the protein-design workflow. | Settings, status, structures, scores, and downloads in one project | Organized through your own tooling | Notebook and file-session dependent | Platform-dependent |
| Billing modelUnderstand what the design campaign costs. | Pay only for the GPU time used by the project | GPU, storage, and infrastructure costs | Free and paid usage tiers | Often plan- or contract-based |
Choose the scientific method that fits the design objective.
Translate a target and design intent into a reproducible campaign.
Move from prepared inputs to active design work.
Compress campaign time by running design work across multiple GPUs.
Stay close to every stage of the protein-design workflow.
Understand what the design campaign costs.
Start exploring
Explore the platform for free. Pay only for the GPU time used by the designs you run.