What is the Hermes Agent?

Hermes Agent is an open-source AI agent developed by Nous Research, engineered to execute tasks, retain learned knowledge, and continuously evolve. By leveraging tools such as terminal access, browser navigation, file system operations, web searches, and code execution, it performs substantive work rather than merely generating textual responses.

What defines an AI agent?

AI agents go beyond simple text generation. When provided with a goal, an agent can independently determine the necessary actions, utilize available tools, analyze outcomes, and persist in working toward the objective.

While many AI agents can now handle multi-step tasks using various tools, the critical distinction lies in post-task behavior. Some agents treat every task as a completely new challenge. In contrast, Hermes is designed to retain valuable knowledge and procedures, allowing its utility to grow over time. For insights into how this applies to running agents on DaDesktop, refer to Agentic AI.

What distinguishes Hermes?

Hermes features an integrated learning loop. It generates skills from experience, refines these skills during operation, retains critical information, and can retrieve data from previous sessions when historical context is required.

Why choose Hermes Agent?

Hermes is particularly beneficial for tasks involving multiple steps, tools, or applications. Rather than manually switching between AI chats, terminals, browsers, and file managers, you can delegate the overarching task to Hermes and allow it to coordinate the workflow.

• Execute complete tasks: Provide Hermes with high-level objectives rather than granular, step-by-step instructions.
• Maintain long-term context: Preserve useful information across separate sessions.
• Create reusable workflows: Store successful methods as skills to avoid re-solving identical problems.
• Automate routine work: Schedule tasks to execute automatically in the background.
• Delegate responsibilities: Utilize distinct subagents for different components of larger tasks.
• Select your model: Connect Hermes to local models or supported cloud model providers.

Capabilities of Hermes

• Software development: Analyze code, troubleshoot issues, implement changes, execute tests, and manage repositories.
• Research: Perform web searches, examine sources, gather data, and synthesize findings.
• Automation: Execute scripts and schedule recurring routines such as report generation and monitoring.
• System administration: Utilize the terminal, install software, configure systems, and resolve technical issues.
• Browser operations: Navigate websites, interact with page elements, and extract information.
• File management: Read, create, edit, search, and organize files.
• Tool integrations: Connect additional services and tools via MCP.

How Hermes improves through usage

Consider a complex task requiring multiple steps. As Hermes solves it, it may identify an effective workflow. Instead of losing this knowledge upon task completion, it can save the workflow as a skill.

When a similar task arises in the future, Hermes can leverage this saved skill as a foundation. If the workflow requires refinement, the skill can be updated. Additionally, its memory retains key facts and lessons learned from previous operations.

Memory and skills serve distinct functions. Memory archives useful information and context, while skills store procedural steps for executing tasks. Combined, they enable Hermes to accumulate knowledge about both its factual understanding and its operational efficiency.

Background work automation

Hermes includes a built-in scheduler for recurring tasks. You can define schedules using natural language, enabling Hermes to perform work independently without requiring an active user session.

This functionality is ideal for reports, monitoring, backups, research, and other routine procedures. Scheduled tasks can also maintain continuity between executions, allowing them to build upon prior results.

Running local or cloud models

Hermes is not restricted to a single LLM. You can connect it to local models or supported cloud providers and switch models without altering the agent's tools, memory, or established workflows.

Using a local model offers greater control over hardware, model selection, and data processing locations. It also allows Hermes to interact with local files and resources without transmitting them to external model providers.

Installing Hermes

Hermes is compatible with Linux, macOS, Windows, and WSL2. You may install the desktop application or utilize the command-line version independently of the desktop interface.

On Linux, macOS, or WSL2, open a terminal and execute the following command:

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

On Windows, run the installer via PowerShell:

irm https://hermes-agent.nousresearch.com/install.ps1 | iex

Post-installation, launch the setup wizard to configure your model provider and other settings:

hermes setup

For detailed installation options and configuration guidelines, consult the Hermes documentation.

Running Hermes on DaDesktop

Hermes requires an environment capable of executing its tools and storing its data. DaDesktop offers a cloud desktop equipped with the necessary CPU, RAM, storage, and GPU resources to run Hermes alongside a local LLM.

This setup is advantageous when the desired model demands more VRAM than is available on your local machine. Select a GPU with sufficient VRAM, launch a DaDesktop environment, install Hermes, and run the model locally alongside the agent.

You can also keep the desktop active for scheduled tasks and long-running workflows, eliminating the need to keep your personal computer powered on.

Learn more about running local LLMs on DaDesktop.