Hermes Agent - Build, Deploy & Scale Autonomous AI

Hermes Agent - Private, GPU-Ready AI Automation on Azure

Get a ready-to-use AI assistant that completes tasks, remembers useful information and automates everyday work. Use the included private local model, run a more powerful model on a GPU, or connect OpenAI, Anthropic or another popular model of your choice.

Product overview

Hermes Agent is an AI assistant that can complete multi-step tasks, use tools, work with files and remember useful information. This ready-to-use Azure VM keeps Hermes separate from your personal computer and includes a local AI model for greater privacy.

You can also use a larger local model on a compatible GPU or connect OpenAI, Anthropic or another popular model of your choice.

Highlights

• Private local-model option: process prompts on the VM without a third-party LLM API; enabled tools and integrations may still share data externally.
• Ready-to-run Hermes Agent with browser dashboard and CLI
• Ollama installed with the latest best-performing local model configured by default
• NVIDIA GPU support for compatible Azure GPU VMs
• Connect OpenAI, Anthropic or another supported model provider
• Pull another compatible open-source model through Ollama

What is included

• Hermes Agent browser dashboard and CLI
• Ollama installed and ready to use
• gemma4:e2b configured as the default local model
• NVIDIA GPU software support for compatible Azure GPU VMs

Run a larger local model on NVIDIA GPU

Choose a compatible NVIDIA GPU VM size, download a model supported by the selected GPU and available memory, then configure Hermes to use that Ollama model:

ollama pull <model_name>

Why use this VM?

• Start faster: Hermes, Ollama, the dashboard and CLI are already installed.
• Keep work separate: the Azure VM does not automatically access files, email or credentials on your personal computer.
• Choose how AI runs: use the included local model, a larger GPU-powered model or a supported external provider.
• Scale when needed: move from a CPU VM to a compatible NVIDIA GPU VM for more demanding local models.

Default model: Latest best-performing local model through Ollama
Access: Hermes browser dashboard and CLI
Billing: Software charge plus separate Azure infrastructure charges

Getting startedCreate the VM, retrieve the Hermes dashboard password using the TechLatest guide, and open the dashboard over HTTPS.

Azure plan pricing, supported regions, GPU availability, VM recommendations, security-type requirements, support terms and exact component versions must be confirmed in Microsoft Partner Center before publication. Model compatibility and performance depend on the selected VM and model requirements. VM separation reduces access to personal-device data but is not a guarantee against data transmission through user-enabled integrations.

Disclaimer: Other trademarks and trade names may be used in this document to refer to either the entities claiming the marks and/or names or their products and are the property of their respective owners. We disclaim proprietary interest in the marks and names of others.