Compare plans & features

Fully-managed Kubernetes clusters provisioned on your own cloud account.

Separate workspaces to group projects and teams under one account.

Isolated environments such as development, staging and production for your apps.

Dedicated virtual private networks that keep your cluster traffic isolated and secure.

Connect your own AWS and other cloud accounts to deploy infrastructure into.

Bring your own domains with automatic DNS management and free TLS certificates.

Invite teammates to collaborate on clusters, apps and deployments in your account.

Deploy container, Git or Helm-based applications on your clusters.

Run multiple independent instances of an application, each with its own configuration and URL.

The number of build and deploy runs included each month across all your apps.

Automate cluster actions on a schedule or trigger — for example scale, power on and power off.

Deploy a single environment across several clusters for higher availability and reach.

Share an environment with another account so teams can collaborate on the same clusters and deployments.

Add and fine-tune worker node groups — choose instance types and scaling to fit your workloads.

Build your own private templates from Helm charts and Docker images: an input form, generated secrets, install stages and outputs, deployable to any environment you allow.

A searchable history of every action taken in your account for security and compliance.

Chat with your documents: a Qdrant vector database, an answering API on Amazon Bedrock with citations, a private web UI and automatic sync from an S3 bucket.

One OpenAI-compatible address for every approved model (LiteLLM on Amazon Bedrock), with keys and budgets per team, an input guardrail, call logs in your S3 bucket and a usage dashboard.

A multi-user JupyterHub with persistent home folders, idle shutdown, keyless S3 access and a one-click connection to MLflow.

Experiment tracking, a model registry and LLM traces, on PostgreSQL with artifacts in S3. Notebooks, the RAG stack and the LLM gateway log to it.

Scores a RAG stack or your own AI app on golden questions (contains, latency and LLM-judge scorers), records every run in MLflow and can score live traffic.

A new deployment is scored against the live one before it is promoted to main; Block stops a worse release, Warn asks you to confirm.

Try a new chunk size, embedding model or top-k next to the live RAG stack, see it scored, then adopt, discard, or roll back within 23 hours. Enabled per account.

Write and compare versions of a RAG stack's prompt, kept in your MLflow stack, and release one only after it is scored against the live prompt. Judge rubrics can be pinned to a version. Part of evaluated settings changes, enabled per account.

One sign-in page with your identity provider (GitHub, Google or any OIDC) in front of MLflow, notebooks and other internal tools.

PostgreSQL or MySQL with nightly backups to a new or existing S3 bucket, with optional KMS encryption.

Valkey or Redis reachable from the cluster, your VPC or the internet over TLS, with nightly backups to S3.

Core networking essentials — ingress (nginx) and automatic TLS certificate management.

Continuous delivery straight from your Git repositories with Argo CD.

Build container images and run continuous-integration pipelines in your cluster.

Cloud-native CI/CD pipelines powered by Tekton.

Collect metrics and trigger alerts on the health of your workloads with Prometheus. Enterprise plan.

Live resource metrics that power autoscaling and real-time monitoring.

S3-compatible object storage running inside your cluster.

Distributed, multi-node MinIO for resilient production-grade object storage.

Automatically manage DNS records through Cloudflare or AWS Route 53 as you deploy.

Durable block-storage volumes backed by Amazon EBS for stateful applications.

Automatically rebalances workloads across nodes for better efficiency and cost.

Package and run your own custom add-ons alongside the built-in catalogue.

Connect GitHub to automatically build and deploy your applications on every push.

Schedule deployments to run automatically at the times that suit your team.

Fine-grained roles, groups and policies that control exactly what each teammate can do.

Lock environment variables and secrets so they cannot be changed without elevated approval — for stricter governance and compliance.

Programmatic access to the platform for scripting and automation.

Give customers a shareable booking page to schedule meetings with you, with custom availability rules and an appointments dashboard.

Save 11% when you choose annual billing over monthly.

How you reach us and how quickly we respond.

A contractual uptime guarantee backed by service credits.

Bespoke integrations built together with our team for your specific workflow.