Player profile
Google AX
An Apache-2.0 declarative orchestrator from Google for running sandboxed agent tasks at scale on Kubernetes, described with kubectl-style Task, Workspace and Model manifests.
Profile
Where it fits
AX schedules each Task as a sandboxed actor on Agent Substrate. A Workspace defines the git repos, MCP servers and tools an agent gets, and a Model sets which LLM the platform uses, with credentials from a Kubernetes secret.
The ax CLI mirrors kubectl with apply, get, describe and delete, plus agent verbs: suspend and resume checkpoint and restore a task, and ax ssh opens a shell inside a running sandbox. A PostgreSQL-backed control plane runs in the cluster.
Builder web analytics
Measure outcomes from AX-run agent work
AX runs and schedules agent tasks. Separately configured product analytics can show the observed user-facing results of a change those tasks helped ship.
This is an optional measurement workflow, not a native AX integration with Agent Analytics. Configure product tracking and authorized analytics access separately; the listing does not install tracking or report AX task telemetry.
First loop to measure
- a team runs agent tasks on AX to prepare a product change
- the reviewed change is deployed to a separately instrumented project
- a human or an authorized agent reads observed product outcomes
- the team uses that evidence to choose the next task
Copyable prompt
If product tracking and analytics access are already configured, inspect the deployed change over an appropriate observation window. Report measured outcomes, missing evidence and limitations. Do not infer causation from before-and-after movement or treat agent task activity as product usage.