[docs-scanner] Model Runner replica count explanation contradicts deployment behavior

[docker-agent]

File: content/manuals/compose/bridge/use-model-runner.md

Issue

The page contains contradictory information about what happens when modelRunner.enabled is false. At the end of the page, it states:

When false, the replica count is 0, and no Docker Model Runner resources are deployed.

But earlier in the same page, under "Docker Desktop", it says:

When modelRunner.enabled is false, Compose Bridge configures your workloads to connect to Docker Model Runner running on the host:

http://host.docker.internal:12434/engines/v1/

Why this matters

A reader following this documentation would be confused about whether Model Runner is running or not when enabled is false. The first statement says it connects to Model Runner on the host (implying it's running), while the second says "no Docker Model Runner resources are deployed" with replica count 0.

The issue is that "replica count is 0" refers specifically to the Kubernetes deployment, not to whether Model Runner is available. The host-based Model Runner (on Docker Desktop) is separate from the Kubernetes deployment.

Suggested fix

Clarify that the replica count refers only to the Kubernetes deployment, not to Model Runner availability overall. For example:

"When false, the Kubernetes deployment replica count is 0 (no in-cluster Model Runner is deployed), and workloads instead connect to the Model Runner instance running on the Docker Desktop host."

---

Found by nightly documentation quality scanner