An orchestrator for what you have, or a platform that brings it
Humanitec’s Platform Orchestrator is a backend for an internal developer platform. A developer, or an AI agent, describes what a workload needs, and the Orchestrator matches that request against rules your platform team wrote, then provisions and deploys it through your own infrastructure code. Its homepage pitch in 2026 is “Let AI build. On your terms.”: agents provision infrastructure within the rules the platform team defines. Humanitec says there is no need to change your existing CI/CD setup and that it integrates with your existing infrastructure code; it connects them rather than replacing them. It also offers its own portal, and has long described portals such as Backstage as a complement to the Orchestrator.
Ankra starts from the other end. It builds the Kubernetes cluster on the provider you choose, runs CI and delivery onto it, and keeps operating it with security ranking, cost estimates, backups and an AI assistant. Where Humanitec expects you to bring the modules that create a cluster or a database, Ankra ships that work as part of the product.
Humanitec suits a platform team with an existing estate of infrastructure code that wants one rules engine in front of it for hundreds of developers and agents. Ankra suits a team that wants the platform itself, built and run for Kubernetes.
Where the infrastructure comes from
In Humanitec, the Orchestrator creates and updates a fleet of clusters based on rules provided by the platform team, reusing templates in Terraform, Pulumi or Crossplane. Databases, queues and DNS work the same way: a workload asks for a resource, and a rule picks the module that provides it. The Orchestrator targets EKS, GKE, AKS, serverless ECS and VMs, with reference architectures for AWS, Azure, Google Cloud and OpenShift. Since August 2026, built-in drivers let provisioning run entirely inside your own network.
Ankra builds kubeadm or k3s clusters on Hetzner, OVHcloud, UpCloud, DigitalOcean, AWS EC2 and Proxmox VE, creates EKS, GKE, AKS, DOKS, OVHcloud MKS and UpCloud UKS, and imports any conformant cluster. On clusters it builds, it handles upgrades, node groups and power schedules, node autoscaling on the five public clouds and rolling OS security patching. Add-ons go in as stacks of Helm charts in dependency order. Ankra’s managed services, including PostgreSQL and Valkey, run on your own cluster and are in closed beta. If you manage infrastructure as code, Ankra’s Terraform provider covers what the CLI can create.
The trade-off is scope. Humanitec reaches beyond Kubernetes to ECS and VMs and works with whatever infrastructure code you have. Ankra is Kubernetes-first and asks you to write less of that code.
Building and releasing
Humanitec does not build images or run CI. It plugs into your existing pipeline and takes over from the artefact: deploying by rule to each environment, with progressive rollouts, rollback, drift detection, impact analysis and ephemeral environments. In June 2026 it added incremental deployments that re-provision only what changed. Developers can describe workloads with Score, an open workload specification that is a CNCF Sandbox project.
Ankra Pipelines run tests and rootless builds in your own cluster, with Semgrep, Checkov and Trivy and a policy gate that decides which image may be published. With deploy provenance turned on, an application refuses images the gate did not sign, and a merge train on GitHub tests each pull request as the exact commit it will merge. Applications generate the build files, or keep the CI your repository already has, and deploy every successful build. Applications work with GitHub repositories, and some documented pipeline stage kinds, including in-pipeline approval, do not run yet.
Security, cost and compliance
Humanitec’s security page states SOC 2 Type II certification, and its SaaS runs on Google Cloud and AWS. Its pricing puts RBAC on Pro and SAML and audit logs on Enterprise. For a buyer whose procurement needs a SOC 2 report today, that matters.
Ankra’s Security Center lists vulnerabilities in running workloads across every cluster and ranks them by CISA’s Known Exploited Vulnerabilities catalogue and FIRST’s EPSS scores. Cloud Cost estimates spend from list prices per namespace, stack and project, and backups write to a bucket you own. The audit log is kept for 365 days on every plan, and single sign-on is OIDC on Enterprise, with no SAML. Ankra’s control plane runs in the EU, and its certification status is on the trust page.
How the AI works
Humanitec’s AI story is governance for agents you bring. The Orchestrator sits between what an agent requests and what gets provisioned: agents get least-privilege service users scoped per project and environment, with dry-run and plan-only modes, and requests that fit policy are provisioned without a human step. Its MCP server gives coding tools such as Cursor and Claude Code product guidance on the Orchestrator’s concepts; Humanitec describes it as a knowledge source, not as a way to act on the Orchestrator.
Ankra’s AI operates the platform itself. AI Insights analyses every firing alert and can also run scheduled health scans. The AI Gateway reviews pull requests on GitHub, GitLab and Bitbucket Cloud, and on GitHub Ankra AI can fix a failing pull request on its own branch. In-product AI writes wait for a person by default, under autonomy controls. Ankra also has an MCP server with 500+ tools; its writes are audited, and the MCP client decides how to confirm them.
When to choose Ankra, and when to choose Humanitec
Choose Ankra when you do not have, or do not want to write, the modules that create clusters, databases and add-ons, and you want CI, delivery, security ranking, cost and backups for Kubernetes in one product.
Choose Humanitec when
- you have a platform team and an established estate of Terraform or OpenTofu modules, CI and clouds, and want one rules engine in front of it
- your workloads span more than Kubernetes, such as ECS and VMs
- you want a governance layer for many teams’ AI agents across that estate
- you need a vendor with a SOC 2 Type II report, SAML single sign-on or a self-hosted, air-gapped orchestrator
Compare the work that remains
- Create the foundation. Note who writes and maintains the code that creates a cluster and a database.
- Ship a release. Include tests, scans, an approval and a rollback, and note which tool runs each step.
- Add a platform component. Install an operator or monitoring stack and note how it is versioned across environments.
- Investigate a vulnerability. Find every affected running workload.
- Compare total cost. Humanitec’s Teams plan is €1,999 a month with 5 users, and Pro is €4,999 a month with 50 users, after a free trial. Ankra’s free tier includes 30 worker vCPU; beyond that Startup costs €0.01 per worker vCPU-hour (about €7.20 a month). Both sit on top of your cloud or server bill, and Humanitec also needs the CI and infrastructure code it orchestrates.
See Humanitec’s pricing and Ankra’s pricing.
Reviewed 9 November 2026 against Humanitec’s documentation and pricing. Product scope and plans can change.