Executor documentation
The Executor packages any application, its portable language runtime and its security policy into one standalone, signed executable, and runs it under every restriction you configure: sandboxing, CPU and memory caps, exit events and auto-scaling. The Executor's command-line tool is oee.
The Executor is in early preview — message us for access.
Where to start
- New to the Executor: read the Overview and How it works.
- Ready to try it: Install, then build your first binary in minutes.
- Looking something up: the CLI reference and the
oee.yamlreference. - Something went wrong: Troubleshooting and FAQ.
The sections
Overview
What the Executor is, how a wrapped binary runs, and when it is the right fit.
Getting started
Installing the
oeecommand-line tool and building your first single binary.Concepts
The single-binary layout and its five segments, the core and extended supervisor variants, and the runtime cache.
Guides
Runtimes and adapters, the guided configuration flow, sandboxing, CPU and memory capping, exit events, stopping and exit codes, auto-scaling, containers and Kubernetes.
Reference
Every
oeesubcommand and flag, and every section and key ofoee.yaml.Security and integrity
Signing, BLAKE3 and Ed25519 verification, secret scrubbing and the edge cases the Executor handles.
Performance specifications
The design targets for binary overhead, startup, supervisor footprint, stop latency, cap accuracy and load balancing.
Troubleshooting and FAQ
Startup refusals, exit codes, container restrictions and common questions.
One binary, every restriction
You bring a built application, choose its runtime, give its start command, and configure exit events, capping, auto-scaling and the sandbox. oee build returns one binary file. That binary verifies its own integrity, applies your policy and starts your application, on Linux, Windows, macOS or in an OCI container, with no daemon, orchestrator or interpreter installed on the host.