Roadmap
What's next after v0.1.0 — OpenTelemetry export, multi-modal inference, a model-driven loop, multi-app orchestration, and an optional managed cluster layer.
v0.1.0 is the runtime core. These are the investments that follow. They are directional, not dated.
Already shipped since this page was first written
Listed here because they used to be on the list below, and a roadmap that never removes anything is a wish list:
- The TypeScript & Python SDKs — both ship over the gRPC gateway.
- A web console —
kx serveserves it at127.0.0.1:8888. - A Prometheus
/metricsendpoint — opt-in, RED metrics.
On the way
- OpenTelemetry export — wire runtime telemetry into your existing observability stack. The opt-in Prometheus endpoint ships today; OTel does not.
- Multi-app orchestration — one app invoking another. Today apps chain inside one app, and travel between machines as a file you hand over; app-to-app invocation is not a thing the runtime does yet.
- Idle suspend — a served app that stops holding a process when nobody is using it. Today a hosted app runs until you stop it.
- Audio multi-modal inference — extend local inference beyond text and images.
- A model-driven dynamic loop — the model computes the topology decision (how to fan out / recover), instead of a fixed recipe shape.
- An opt-in cluster layer for managed multi-node — the hosted side of the coordinator/worker model.
Phases
The runtime (phase 1) is what ships today. Platform & orchestration — multi-app workflow definitions, a run-graph compiler, and managed multi-node — is the next sequenced phase.
Built in the open
Kortecx runs as a managed service — sign-up is open, from the button in the header of any page here; enterprise and services are contact-sales. A public runtime release is planned under a separate repository.