The console, labelled LOCAL.
cloudburrow up serves it at http://127.0.0.1:9090. It is a view, not a second
system: everything it shows is read through the same APIs your SDK calls, and a control appears only
where the backend can perform it.
Product icons are hidden in these captures.
Captured 2026-09-28 from a console built at 5da2bd0. Resource names are examples. The captures show that build's screens.
Inspect the resources your tests use.
Start with the dashboard, then go straight to a service, the request logs or the instance controls.
- Local address · 127.0.0.1:9090
- Data source · The same APIs as your SDK
- Capture revision · 5da2bd0
Start with the local instance.
Open a service to inspect its resources, or go to the instance controls to manage the test environment.
Go straight to the detail you need.
Storage
Upload, download, preview, copy, move and compose objects.
Pub/Sub
Publish and pull.
Cloud Run
Deploy, edit and delete services; revisions; jobs and their executions.
Data
SQL editors for Cloud SQL and Spanner that read by default, query builders for Firestore, Datastore and Bigtable, BigQuery preview.
Kubernetes
Read-only workloads, pods, services, jobs, events, nodes and storage, with ownership per object and metric history.
Logs
A live Logs Explorer over server-sent events, with pause, reconnect replay and credential redaction.
Instance
Save and load state, reset, seed, and inject faults.
Terminal
Activate terminal opens a drawer, like Cloud Shell, with a shell in a pod in the local cluster, never on your machine. It has gcloud pointed at this instance and read-only kubectl, follows the toolbar’s project, and keeps up to eight tabs, each its own session that survives a reload.
No host paths or credentials are mounted, and gcloud runs with credentials disabled. Its connection requires a loopback peer and the console’s own origin. Unlike Cloud Shell, it has no Terraform, editor, web preview or persistent home directory. The terminal is new in v0.3.0; v0.1.0 has none.
Console additions in v0.3.0
These screens are new since v0.1.0. They may not appear in the dated captures above.
- Cloud Run
- Manage revision traffic and tags, and set service scaling and its CPU and concurrency targets when you deploy or edit a service. Inspect jobs, their executions and tasks, and use the Worker pools and Instances views.
- Cloud Storage
- Inspect object generations, restore soft-deleted objects and buckets, manage folders, and configure bucket notifications.
- Memorystore
- Inspect Redis instances and clusters or Valkey instances when the corresponding service is enabled. The forms call the local admin APIs.
- IAM & Admin
- List the project’s and the organization’s custom roles beside Google’s predefined roles, and create, edit, delete and undelete custom roles. Manage service accounts, their keys and their own permissions. Roles made here are the ones gcloud and Terraform see.
- Permissions
- Inspect and edit IAM policies on supported resources, with IAM Conditions where Google allows them, and test their recorded permissions. Most services store those policies without enforcing them.
- BigQuery connections
- A page per BigQuery connection, with its IAM policy. Remote models and object tables use these connections.
- Local AI
- With the configured ONNX build, use Local models to manage downloads and try the Embeddings view. The playground offers the generation options local generation honours.
Themes and responsive
Light and dark follow your system or a choice the console remembers, and the layout holds down to a phone.
What it doesn't do
- Not pixel-identical to Google's console; there are no reference captures to compare against
- No billing or quota screens
- Apart from five named IAM checks, policies do not authorize or refuse emulated service requests
- The Kubernetes views are read-only and do not apply manifests or expose a GKE control plane
Build against Google Cloud APIs, locally
Free and open source under Apache-2.0. No account, no sign-up, no Google Cloud bill.