Skip to content

Hermes Agent Beginner's Guide

Introduction

Hermes Agent is an open-source AI agent created by Nous Research. Unlike a normal chatbot, Hermes can interact with your computer, use tools, remember information, browse the web, and automate tasks.

Official documentation:

  • https://hermes-agent.nousresearch.com/
  • https://www.hermes-ai.net/

What Is Hermes?

Hermes is a personal AI agent that runs on your machine and can:

  • Chat with you in the terminal.
  • Use different AI models.
  • Execute commands.
  • Browse the web.
  • Read and write files.
  • Remember previous interactions.
  • Connect to external tools.
  • Automate workflows.
  • Be customized and reconfigured.

Think of Hermes as a programmable AI assistant rather than just a chatbot.

Installation

Hermes supports:

  • Linux
  • macOS
  • WSL2 on Windows
  • Docker

Native Windows installation is not currently recommended.

Quick Installation

Open a terminal and run:

curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

After installation:

source ~/.bashrc

or:

source ~/.zshrc

Start Hermes:

hermes

Alternative Installation Using pipx

Install pipx:

python -m pip install --user pipx
python -m pipx ensurepath

Then install Hermes:

pipx install hermes-agent

Docker Installation

docker run --rm -it \
-v ~/.hermes:/root/.hermes \
-e OPENAI_API_KEY=$OPENAI_API_KEY \
nousresearch/hermes-agent:latest

Verify Installation

Check the version:

hermes --version

Run diagnostics:

hermes doctor

Hermes will report missing dependencies or configuration problems.

Initial Setup

Run:

hermes setup

This setup wizard allows you to:

  • Choose an AI provider.
  • Configure API keys.
  • Enable tools.
  • Configure memory.
  • Customize your environment.

Choosing a Model

Run:

hermes model

You can connect Hermes to:

  • OpenAI
  • Anthropic
  • Google Gemini
  • OpenRouter
  • Ollama
  • LM Studio
  • Any OpenAI-compatible endpoint

Basic Commands

Start chatting:

hermes

Run setup again:

hermes setup

Change the model:

hermes model

Configure tools:

hermes tools

Set individual configuration values:

hermes config set

Run diagnostics:

hermes doctor

Reconfiguring Hermes

One of Hermes' biggest strengths is that you can reconfigure almost everything.

Change the AI model

hermes model

Example:

  • GPT-4
  • Claude
  • Gemma
  • Qwen
  • DeepSeek
  • Local Ollama models

Configure Tools

hermes tools

You can enable or disable:

  • Web search
  • Browser automation
  • File access
  • Memory
  • External integrations

Change the Execution Backend

Hermes supports several backends.

Local

Runs directly on your machine.

hermes config set terminal.backend local

Docker

Provides isolation and safety.

hermes config set terminal.backend docker

SSH

Execute commands on another machine.

hermes config set terminal.backend ssh

Use serverless GPU resources.

hermes config set terminal.backend modal

Using Local Models with Ollama

Install Ollama:

curl -fsSL https://ollama.com/install.sh | sh

Download a model:

ollama pull qwen3

Start Ollama:

ollama serve

Then configure Hermes to use Ollama:

hermes model

Choose:

Ollama

Now Hermes runs using a local model.

Memory

Hermes can remember:

  • Conversations.
  • Preferences.
  • Files.
  • Previous work.

This allows long-running workflows and persistent assistants.

Skills

Hermes supports skills that extend its abilities.

Examples:

  • Coding assistants.
  • Research workflows.
  • Project management.
  • Document generation.
  • Task automation.

Practical Things You Can Do

Software Development

Ask Hermes to:

  • Write code.
  • Explain bugs.
  • Refactor projects.
  • Generate tests.

Research

Hermes can:

  • Search the web.
  • Summarize information.
  • Create reports.
  • Compare technologies.

File Management

Hermes can:

  • Create files.
  • Edit files.
  • Organize folders.
  • Analyze documents.

Automation

Examples:

  • Daily reports.
  • Monitoring projects.
  • Email workflows.
  • Data processing.
  • Scheduling tasks.

Updating Hermes

Update to the latest version:

pipx upgrade hermes-agent

or reinstall using the installer.

Troubleshooting

Command Not Found

Reload your shell:

source ~/.bashrc

or

source ~/.zshrc

Check Configuration

hermes doctor

Start Setup Again

hermes setup

Learning Path

Beginner

Learn:

  • Installation
  • Setup
  • Basic chat
  • Models

Intermediate

Learn:

  • Tools
  • Memory
  • Skills
  • Configuration

Advanced

Learn:

  • Ollama integration
  • Local models
  • Docker backend
  • SSH execution
  • Custom skills
  • Automation workflows
  • Multi-agent systems

Conclusion

Hermes is more than a chatbot. It is an AI agent platform that can be customized and extended to become a powerful assistant for coding, research, automation, and daily work.

As you become more comfortable with Hermes, you can gradually move from simple conversations to advanced workflows involving local models, memory, tools, and automation.