Product · Orchestration & Runtime · Production
Lightweight container orchestration delivered as a managed service. Sign up, connect your nodes with a one-liner, and start deploying. Web dashboard. Autoscaling. Canary deployments. AI-powered operations. No platform team required.
The problem
Plain container compose works well for small deployments — until it doesn't. Once you need autoscaling, canary deployments, secret management, multi-network attachment, or proper RBAC, the conventional jump is to a heavyweight orchestrator: a dedicated control plane, a state store cluster, an ingress controller, multiple operators, charts, manifest sprawl, and the platform engineer to run it. The complexity tax is real and the resource overhead is real.
Most operations teams are not running thousands of nodes. They are running 50–100 workloads on a handful of hosts and they want orchestration that matches that scale. Odysseus is the orchestrator built for them: a single binary, a small dashboard, and the operational features that matter — without the multi-component cluster you don't need.
What it does
Single-binary control plane
The control plane runs on a single host. The agent is embedded in the same binary. No separate scheduler, controller manager, ingress controller, or secret operator to deploy and maintain.
Metric-based autoscaling
Define the target metric and bounds; Odysseus scales the workload between them. No per-workload custom resources, no controller installation, no cluster-wide configuration.
Canary deployments
A new revision rolls out to a fraction of the traffic, error and latency are measured against the baseline, and the rollout either promotes itself or rolls back. No service-mesh dependency.
Multi-network support
The capability that lets a workload sit between two perimeters cleanly — one network for the public side, another for the back-end systems it integrates with.
Secret management
The right default for credentials and keys. Process listings, crash dumps, and `docker inspect` no longer leak secrets.
Web dashboard + RBAC
Admin, operator, developer, and read-only roles. The dashboard is the daily driver; the CLI exists for scripted work. No external dashboard tool required.
Where it fits
The orchestration core is operated for you — patched, monitored, kept available. Your nodes enrol via a one-liner and an encrypted mesh, and your workloads run on your own hosts. The customer relationship with the data stays where it should: with you.
Mid-size organizations with one or two operators. The right scale to need orchestration, the wrong scale to dedicate a platform team to running one. Odysseus produces the operational outcomes — autoscaling, canary, secrets, RBAC — without the headcount.
For sovereign buyers, regulated environments, and air-gap deployments, the same control plane is available to run inside your perimeter on the Enterprise tier. The single-binary control plane runs on a single host with under 150 MB of RAM — a regional operator's edge site, a Northern anchor institution, a field office. We support it under SLA when you want us to operate it for you.
Maria, Kacha, ewii, IROS, and Musa all run on Odysseus inside customer environments. Customers running Odysseus can adopt any of the rest with a single command — the orchestration substrate is already there.
Patterns that use this product
Sign up · connect a node · deploy
Sign up for the managed control plane, run a one-liner to enrol your first node, and start deploying. Or talk to an engineer about an evaluation, a self-hosted Enterprise control plane, or a sovereign edge build.
