Google Cloud · Cloud service
Google Compute Engine with OrchestrAI
Catalog exported 2026-09-02
Operate Compute Engine from chat: VMs, instance templates, managed instance groups, autoscalers, Cloud Armor.
OrchestrAI exposes 7 Compute Engine operations: 2 are low-risk (read-only or low-impact), and 5 create or modify resources and run only after you confirm the plan. 1 of them also carries a step-level approval gate.
What teams use it for
Platform engineers use OrchestrAI to build the full autoscaling chain on Compute Engine: an instance template, a managed instance group from it, and an autoscaler with a target utilization, described in a few sentences instead of a long gcloud session. One-off VMs for testing and a Cloud Armor policy for the public edge are the other routine requests. You cannot stop, start, or resize a running VM from chat yet, and there is no operation to delete a managed instance group, so those steps remain in the console.
Every Compute Engine operation, with its risk level
| Operation | What it does | Risk | Step-level approval |
|---|---|---|---|
Create GCP Cloud Armor Security Policy |
Create a GCP Cloud Armor security policy for DDoS and WAF protection | Low risk | No |
List Compute Instances |
List GCP Compute Engine VM instances | Low risk | No |
Create Autoscaler |
Create an autoscaler for a managed instance group | Creates resources | No |
Create Compute Instance |
Create a GCP Compute Engine VM instance | Creates resources | No |
Create Instance Template |
Create a GCP instance template for managed instance groups | Creates resources | No |
Create Managed Instance Group |
Create a GCP managed instance group (MIG) | Creates resources | No |
Delete Compute Instance |
Delete a GCP Compute Engine VM instance | Modifies existing | Yes |
Risk tiers come from the catalog: low is read-only or low-impact, medium creates resources and is reversible, high modifies existing resources, destructive may lose data. Every plan that creates or changes resources is shown with its cost estimate and waits for your confirmation. Operations marked with a step-level approval pause again on their own step. Destructive operations require a typed risk phrase.
Prompts that work
- List all Compute Engine instances in us-central1-a and show their machine types
- Create an instance template called web-v3 with e2-medium and the debian-12 image, then a managed instance group web-mig with 3 instances in us-east1-b
- Add an autoscaler to web-mig that keeps CPU at 60 percent, min 2, max 10
Before anything runs
Every mutation shows its plan, cost estimate, and blast radius, then waits for your confirmation. Destructive operations require a typed risk phrase. Credentials are minted per run through OIDC federation and discarded afterward; nothing you create here is invisible later, because every resource lands in the desired-state ledger where drift is detected and can be converged. Details on the security page.
Frequently asked questions
- Can OrchestrAI delete a Compute Engine VM?
- Yes, with a confirmation step. Deleting an instance is rated high risk, so OrchestrAI presents the plan and waits for you to approve it before the VM is removed.
- Does OrchestrAI support Cloud Armor?
- It can create a Cloud Armor security policy for DDoS and WAF protection, which is a low-risk operation. Attaching that policy to an existing backend service is not yet an available operation.
- Which Compute Engine operations need an extra approval step?
- One operation carries a step-level approval gate on top of plan confirmation: Delete Compute Instance. None of them is classed destructive.
Other Google Cloud services
Related integrations
Try it on your own account
Connect your cloud read-only and see your resources, drift, and costs before anything runs. $5 minimum to start. Unused credits refunded in your first 14 days.
Unused credits refunded in your first 14 days.