AGENTIC CLOUD PLATFORM

Less infrastructure.More discovery.

Supercloud is an agentic cloud platform. Pipelines, monitoring, AI assistant in one place, tuned to the way your domain works, for you and your agents.

Talk to our team

Supercloud Platform

From running pipelines to understanding results — the assistant analyzes and suggests, you decide

PipelinesMonitoringAI Assistant
Your own steps

Add a Python script or a container image from an allowed registry and run it like any other tool.

Learn moreNew Job form with a custom container image URL from the registry.
Pipelines from templates

Connect tools into a pipeline, fan out over many inputs, and rerun only the step that needs it.

Learn morePipeline detail page with the flow graph from its template.
Ready-made bio tools

Start from MMseqs2, Foldseek, US-align, ProteinMPNN and ESM-2, each ready to run from a form or a REST call.

Learn morePipeline catalog with a ProteinMPNN design tool ready to launch.
New Job form with a custom container image URL from the registry.Pipeline detail page with the flow graph from its template.Pipeline catalog with a ProteinMPNN design tool ready to launch.
Kubernetes clusters

Watch node health, GPU allocation and CPU, memory and network use on each Kubernetes cluster.

Learn moreKubernetes cluster metrics with live CPU utilization.
Slurm clusters

Follow partitions, node states and queue load on each Slurm cluster as jobs come and go.

Learn moreSlurm cluster overview with its operator, login and worker replicas.
Jobs

See running and pending jobs and the resources each one uses, from submission to finish.

Learn moreJobs list with a summary of one job’s duration, CPU cores and memory.
Kubernetes cluster metrics with live CPU utilization.Slurm cluster overview with its operator, login and worker replicas.Jobs list with a summary of one job’s duration, CPU cores and memory.
Metrics and structures

See the scores from a run next to the structures they describe, without opening each output file.

Illustrative Boltz-2 run with its model ranking next to the predicted structure, colored by per-residue pLDDT.
Analysis on request

Ask about a run in plain language and get an answer drawn from its own results and logs.

Illustrative assistant answer, drawn from the run's results, on why a binder design fell below the ipTM cutoff.
Next-step suggestions

Get suggestions for what to run next, such as which designs to refine and which to drop.

Illustrative assistant suggestion to keep one binder design and drop another, with a ProteinMPNN job waiting for approval.
Illustrative Boltz-2 run with its model ranking next to the predicted structure, colored by per-residue pLDDT.Illustrative assistant answer, drawn from the run's results, on why a binder design fell below the ipTM cutoff.Illustrative assistant suggestion to keep one binder design and drop another, with a ProteinMPNN job waiting for approval.

What you can run

Pick the level you want to work at, from a cluster of your own to a single API call.

Who it’s for

Supercloud takes care of everything underneath — failed nodes, restarts, drifting environments —
so teams that share compute can count on their work running the same way every time.

  • Labs and teams with many projects

    Let project teams run clusters and jobs on their own, with OIDC login, invitations and project roles — without an admin in the loop for every request.

  • AI and HPC teams

    Submit containerized jobs to Slurm and keep files, images, metrics and logs in one place, so a failed run is easy to trace and repeat.

  • Domain teams with their own toolchains

    Run established tools and pipelines without installing or maintaining any of them. Bio pipelines come ready to use.

  • Organizations with data rules

    Run on your own infrastructure, with jobs cut off from outside networks by default and the same environment on every node.

Nothing runs without you knowing

Every run is recorded, priced before it starts, and approved by a person.

  • Results you can reproduce

    Each job records the tool definition, image digest, settings, input checksums and resources used.

  • Costs before you run

    See an estimate before you submit a job, then the actual usage in the same unit when it ends.

  • Agents with approval

    Connect an agent over MCP. It can find bio tools and request runs, and work starts only after a person approves.

Job record

Tool definition
proteinmpnn · 3f9a1c07
Image digest
sha256:8b2e6d…41d0
Settings
num_sequences 16, sampling_temp 0.2
Input checksum
sha256:c07a15…9e12
Estimated usage
1.2 CU
Actual usage
1.1 CU
Approved by
Project admin
Silicon to service

Supergate designs and ships its own NPUs and SoCs.Supercloud is the platform we built on them and the services you run on top.