Nous Research unveils a profile builder for Hermes Agent

On June 11, 2026, Nous Research released an update to the Hermes Agent panel: an AI agent profile builder that, for the first time, combines the choice of identity, model, skills and MCP servers into a single step-by-step flow inside a web interface. Before this, setting up an agent required multistep work through the command line (CLI), a noticeable barrier to entry for users without a technical background.

What changed in the interface

Previously, building an agent on the Hermes Agent platform required a sequence of actions in the terminal: the agent's identity was set separately, the model was chosen separately, and skills and MCP servers were connected separately. The new dashboard flow collapses all these steps into one form. The user goes through the configuration once and gets a ready-made agent profile, without switching between commands and settings files.

• Company: Nous Research
• Product: Hermes Agent (panel/dashboard update)
• What's new: a single flow for identity, model, skills and MCP servers
• What it replaces: multistep setup through the CLI
• Announcement date: June 11, 2026 (according to MarkTechPost)

Why an agent needs MCP servers

MCP (Model Context Protocol) is an open protocol that lets AI models connect securely to external tools, databases and services, so that each developer doesn't have to write their own integration from scratch. For agent platforms like Hermes Agent, it is a key component: through MCP servers, an agent gains access to the file system, third-party service APIs or specialized tools, expanding the range of tasks it can perform autonomously. Putting MCP server selection directly into the profile builder means users no longer have to write connection configuration by hand. It becomes part of the wizard, on par with choosing a model.

The other two components of the flow, identity and skills, are not decorative either. Identity defines how the agent presents itself and which persona or system prompt shapes its behavior when it talks to the user and to other services. Skills, in turn, are modular sets of capabilities that can be attached and detached as needed without rewriting the agent's base configuration from scratch, much as plugins extend a program's functionality without changing its core. Combining these three layers (who the agent is, which model drives it and what it can do) in one wizard spares users from remembering in which order and in which file each parameter is configured.

What does this mean for developers?

Nous Research has long been known as the developer of the open Hermes family of models: fine-tunes of large open-source LLMs focused on instruction-following behavior and function calling. The move from "model as product" to "agent as product" is a logical extension of that line. Instead of users assembling an agent around a Hermes model themselves, the company offers a ready-made visual builder. For developers and small teams, this lowers the barrier to building autonomous AI agents: there is no need to know CLI command syntax or the structure of configuration files, and a single flow in the browser is enough.

The trend this update fits into is broader than one product. As agentic AI systems grow in popularity, platform developers are moving one after another from terminal tools to graphical interfaces, lowering the barrier to entry for non-programmers. Agentic systems are gradually becoming a product not only for engineers but for a wider audience that wants to assemble a working assistant quickly, without digging into infrastructure details.

Nous Research itself has historically positioned itself as a notable player in the open-weight segment of AI models rather than a closed corporate lab, and it continues that openness in its agent-building tools. The dashboard that combines identity, model, skills and MCP servers is a logical complement to the Hermes model lineup: users now have not only an open model but also a clear interface to wrap it in a working agent, without sorting out the details of MCP integration or the structure of configuration files by hand.