Run frontier AI models on our GPUs
Your workflow stays local — the heavy steps run on our managed GPUs, powered by frontier open models for structure prediction, docking, and genomics. Register and start on the free tier: no cloud account, no setup.
Free during the Technical Preview · no personal cloud account required
One workflow, mixed execution.
Design and run your workflow on your own machine, exactly as you do today. Flag a heavy node to run on Salpa Compute — only that step goes remote, and its results return into your local graph.
Frontier AI models, ready to run.
Deployed, state-of-the-art open models you can drop into a workflow today.
Boltz-2
Predicts a protein’s 3D structure from its amino-acid sequence (AlphaFold-class open model).
Chai-1
Predicts 3D structures of complexes — proteins together with ligands or nucleic acids.
DiffDock
Blind protein–ligand docking: finds binding poses without a predefined pocket.
Evo2
DNA foundation model — score, embed, or generate genomic sequences.
ProteinMPNN
Inverse folding: designs amino-acid sequences that fold to a given backbone.
New models, in hours.
We can turn a new ML model or tool into a live GPU node in hours, not months — the catalog keeps growing. Tell us what you need.
Request a model →Register and run — no cloud account.
Access is automatic: sign in and the free tier is provisioned for you. No waitlist, no approval.
Register in Salpa
Sign in with Google or email — no personal cloud account, no waitlist.
Free tier, instantly
Your account is provisioned on the free tier automatically on first use.
Run a cloud node
Drop a GPU node into your local workflow; it runs on our managed GPUs.
Results flow back
Outputs return into your local graph — one run, mixed local and remote.
Free while we’re in Technical Preview.
Every account starts with a free monthly GPU allowance — enough to try every node. Your workflow orchestrates locally; only the heavy nodes draw on it.
Need more capacity? contact us
Powered by Modal and Google Cloud.
Salpa Compute runs on production-grade GPUs from Modal and Google Cloud, who also support us through their startup programs. We run the infrastructure so you can focus on the science.
Bring GPU nodes into your workflow.
Start with Salpa, sign in, and send the heavy steps to a managed GPU.