Run Your Infrastructure by Conversation
Verge CLI turns VergeOS, the leading VMware alternative, into an AI-powered Cloud Operating System through three components: the Verge CLI command-line interface, an MCP server built on the open Model Context Protocol, and a set of agent skills. Together they let Claude Code, OpenAI Codex, or a local model you run yourself build networks, deploy workloads, and diagnose faults in plain language. The administrator decides what the assistant runs on its own and what needs approval. Generally available June 23, 2026.
A VMware exit usually stops at the hypervisor
Most teams approach a VMware exit as a licensing problem. They cut the per-core costs, swap the hypervisor, and call the project done. The larger opportunity, an environment that is easier to operate and ready for AI, goes unclaimed. Worse, the day-two complexity that made VMware painful simply moves to the new platform, and the team inherits the same operational burden under a different name.
- Migrations stall when a team is not certain how the new environment behaves.
- Layered stacks expose a separate API per product, so day-two operations stay complex.
- Infrastructure “AI” rarely goes past a chatbot that answers documentation questions.
- Advanced platform features get postponed for months because no one has time to learn them.
Verge CLI makes the platform easier to operate
VergeOS has always been API-first, so Verge CLI was a natural extension. It maps to the full VergeOS API across compute, storage, networking, and data protection. An MCP server built on the open Model Context Protocol connects AI platforms, and a set of agent skills supplies the know-how, so an assistant can act on the environment.
- The assistant reads the environment and proposes the work; the administrator decides what it runs on its own and what needs approval.
- One command set covers compute, storage, networking, and data protection through a single code base.
- An assistant guides the migration and the validation that follows, so teams reach production faster.
- The same assistant teaches advanced features in plain language, so customers use the depth of VergeOS sooner.
What’s In It For You
Verge CLI is built for the person who runs the environment. It removes the busywork and shortens the learning curve, and it keeps you in control the entire time.
Skip the Learning Curve
The assistant guides you through VergeOS in plain language, so you are productive on day one instead of month three.
Operate the Whole Stack
Create VMs, build networks, carve storage, and set data protection in one conversation, without mastering four separate tools.
Stay in Control
You decide what the assistant runs on its own and what waits for your approval. It never acts outside the limits you set.
Find Root Cause Faster
The assistant reasons against VergeOS documentation and traces a fault across compute, storage, and networking, so you fix the real problem, not the symptom.
Reclaim Your Time
Hand off routine work like VM creation, workload moves, and network changes, and spend your hours on the projects that matter.
Grow Your Value
You become the operator of an AI-assisted platform, with deeper reach and more impact, not less.
Three Reasons It Holds Up Against a Fast-Moving Field
The MCP category now includes vendor offerings from Nutanix, Red Hat, and Microsoft, plus community servers for VMware and Proxmox. These pillars keep the focus on what VergeOS does that the others cannot.
One Code Base, One API
A single code base controls compute, storage, networking, and data protection. The assistant operates the whole stack through one API and one tool surface. Rivals expose separate products with separate APIs, which fragments AI control and adds failure points.
Migrate Faster, Adopt Sooner
The assistant guides VMware refugees through the migration and the validation that follows. It then teaches the team advanced features in plain language, so customers reach production faster and use the depth of the platform sooner.
Native, Safe, and Private
The Verge CLI, the MCP server, and the agent skills ship and are supported as part of the product. The MCP server builds on the Model Context Protocol, an open standard, so any compatible client works. The administrator decides what the assistant runs on its own and what needs approval, and local open-weight models keep all data on customer infrastructure.
The Brains, the Hands, and the Know-How
The release is three distinct components: a Verge CLI command-line interface that gives the commands to act, an MCP server that connects AI platforms over the open Model Context Protocol, and a set of agent skills that supply the know-how.
MCP Server
An MCP server built on the open Model Context Protocol gives an AI assistant a secure, scoped view of the environment and its documentation. You decide what the assistant can see.
Verge CLI
One command set maps to the full VergeOS API across compute, storage, networking, and data protection. These commands are the actions the assistant takes.
Agent Skills
A set of agent skills encodes how VergeOS experts design networks, build environments, and run diagnostics, so the assistant works like a seasoned operator, not a generic chatbot.
One Surface Versus Many
With VergeOS the assistant reasons about the whole environment through one tool surface. A request to build a VM, place it on a network, carve storage, and set replication runs as one coherent action against one API. On a layered stack the same request crosses vSphere, vSAN, NSX, and a separate backup product, and each part can half succeed. The single code base removes that class of error.
Root Cause Across the Stack
Through the MCP server, the agent reasons against VergeOS documentation, so its diagnoses come from how the platform actually works rather than a model’s guess. Because one API spans compute, storage, and networking, it traces a fault across the whole stack that tooling stitched across separate products would miss.
From Install to Operating by Conversation
Install Verge CLI
Add Verge CLI to your workstation or management host. It ships and is supported as part of VergeOS, so there is no third-party add-on to maintain.
Connect an MCP Client
Point an MCP-compatible client at the Verge MCP server. Use a cloud assistant such as Claude Code or OpenAI Codex, or a local open-weight model for strict environments.
Verify and Set Limits
Confirm the integration is working, then set what the assistant can see and what it runs on its own versus what needs approval. The administrator sets the limits.
Operate in Plain Language
Create VMs and networks, move workloads, and troubleshoot in natural language. The assistant runs what you have cleared on its own and pauses for approval on the rest.
What the Assistant Does, in Practice
The platform that retires per-core VMware licensing now operates through conversation. The same MCP integration runs against a cloud assistant or a local model, and the choice comes down to where your data is allowed to go.
Cloud frontier assistants
Most organizations are well served by cloud assistants such as Claude Code and OpenAI Codex, which deliver the strongest reasoning and the fastest path to value. Sales leads with this option for the majority of accounts.
Local open-weight models
Teams under strict security or compliance rules run a local open-weight model such as Llama, Qwen, or DeepSeek through a runtime like Ollama. Every operation and all environment data stay on customer hardware. The integration stays the same; only the location of the model and the data changes.
| What the AI works with | VergeOS + Verge CLI |
|---|---|
| What ships | The Verge CLI, an MCP server, and agent skills |
| API the assistant drives | One API across compute, storage, networking, and data protection |
| Cross-domain task | A single coherent operation, not multi-product orchestration |
| Root-cause visibility | One data model across the whole stack |
| Integration standard | Model Context Protocol, an open standard |
| Compatible clients | Claude Code, OpenAI Codex, and any MCP-compatible client |
| Local model option | Llama, Qwen, DeepSeek via a runtime like Ollama |
| Control model | Administrator decides what runs on its own and what needs approval |
| Support model | Vendor-built, shipped with the product |
Five Questions for Any AI-Ops Claim
Use these to compare Verge CLI against vendor and community MCP offerings on the market.
Does one API control the full stack?
Ask whether a single API covers compute, storage, networking, and data protection. On a layered stack an AI agent stitches together several servers, each with its own authentication, data model, and version. VergeOS gives the assistant one tool surface, so a cross-domain task runs as one coherent action.
Is the integration open or proprietary?
Confirm the integration uses an open standard rather than a single vendor’s assistant. The MCP server builds on the Model Context Protocol, so any MCP-compatible client works, including Claude Code and OpenAI Codex.
Who sets the limits on what the AI does?
Check who controls the boundaries. With Verge CLI the administrator decides what the assistant runs on its own, what needs approval, and which actions it can take at all, so it never operates outside the limits you define.
Can it run entirely on your own infrastructure?
For security or compliance requirements, confirm a local model option exists. Verge CLI runs a local open-weight model so all operations and environment data stay in house, with no environment context sent to an outside provider.
Is the capability supported, or a bolt-on?
Determine whether the AI capability is vendor-built and supported or a community add-on. The Verge CLI, the MCP server, and the agent skills ship and are supported as part of VergeOS. Community projects deliver AI access to VMware and Proxmox, but those are unsupported add-ons.
Chat With Your Infrastructure
Type “import this VM and put it on an isolated network” and watch it happen. In this live session, hosted by Truth in IT, we connect an AI agent to a real VergeOS environment and run it through plain language: importing workloads, building a secured three-tier environment, and root-causing a real fault.
- The MCP install, connection, and a check that the integration works
- Virtual machine creation and network creation by conversation
- A three-tier build with firewall rules between tiers, within administrator-set limits
- A system-wide diagnostic that finds root cause from real log lines
- Live Q&A on applying these capabilities in your own environment
- George Crump · CMO, VergeIO
- Jason Yaeger · SVP Product & Engineering, VergeIO
- Larry Ludlow · Chief Architect, Verge CLI, VergeIO
- David Litman · Host, Truth in IT
Resources
VergeIO Launches Verge CLI
The announcement of Verge CLI and the AI-powered VMware alternative, with commentary from VergeIO product and engineering leadership.
Live WebinarChat With Your Infrastructure
A demo-heavy session on June 23 that runs an AI agent against a live VergeOS environment, within the limits the administrator sets.
White PaperAI-Assisted Infrastructure Operations
How a single code base and an open MCP integration let an AI assistant operate compute, storage, networking, and data protection safely on VergeOS.
Frequently Asked Questions
What is Verge CLI?
Verge CLI is the command-line interface in a three-part release for VergeOS: the Verge CLI, an MCP server built on the open Model Context Protocol, and a set of agent skills. Together they let AI assistants operate a VergeOS environment in plain language, with the administrator deciding what the assistant runs on its own and what needs approval.
Which AI assistants does it work with?
The MCP server builds on the Model Context Protocol, an open standard, so any MCP-compatible client works. That includes cloud frontier assistants such as Anthropic Claude Code and OpenAI Codex, and local open-weight models.
Does the AI run my infrastructure on its own?
Only within the limits you set. The administrator decides what the assistant can run on its own, what needs approval, and which actions it can take at all, so it never operates outside the boundaries you define.
Can I keep everything on my own infrastructure?
Yes. Teams with security or compliance requirements run a local open-weight model such as Llama, Qwen, or DeepSeek through a runtime like Ollama, so all operations and environment data stay in house.
Do I still need my VMware administrators?
Yes. The administrator is the buyer and the operator. The assistant guides administrators and shortens the learning curve on a new platform, rather than replacing them.
How is this different from AI on VMware or Proxmox?
Community MCP servers exist for VMware and Proxmox, but those are unsupported add-ons. The Verge CLI, the MCP server, and the agent skills ship and are supported as part of VergeOS, and they control the full stack through one code base and one API rather than a federation of product APIs.
What does Verge CLI cost, and is it available?
Verge CLI reaches general availability on June 23, 2026, to VergeOS customers as part of the platform. For licensing and deployment details, schedule a technical deep dive with the VergeIO team.
See Your Infrastructure Answer Back
A VMware exit should buy more than lower licensing. Verge CLI pairs a single-code-base platform with an AI assistant that guides the migration and the operation, so you move off VMware faster and reach the full value of the platform sooner.
Schedule a Technical Deep Dive
Pick a time for a 1:1 technical whiteboard session with the VergeIO team. Bring your migration and AI questions, and we will walk through Verge CLI against your own environment.