Hermes Agent Tutorial 1: First 5 Minutes — Installation and Quick Win

Tutorial Overview

Series index: Hermes Agent Tutorial Series

This is the first article in the Hermes Agent tutorial series. In just 5 minutes, you’ll have Hermes installed, configured, and responding to your first message.

What you will learn

✅ How to install Hermes Agent on Linux/macOS/WSL2
✅ How to run the setup wizard
✅ How to select your first model
✅ How to verify installation with a test interaction

Why Quick Start Matters

Most AI assistants require lengthy setup before you can try them. Hermes Agent is designed differently: install → setup → first response in under 5 minutes.

flowchart LR
    A[curl install] --> B[hermes setup]
    B --> C[Select model]
    C --> D[hermes]
    D --> E[First response]

    style A fill:#e8f5e9
    style E fill:#fff3e0

Comparison with other agents

Agent	Installation	Setup Time	First Response
Hermes	1 command	2 minutes	< 5 minutes
Claude Code	npm install	3-5 minutes	Immediate
Cursor IDE	Download + config	5-10 minutes	Immediate

Installation

One-command install

Hermes Agent uses a single curl script that handles platform detection and dependency setup:

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

Works on:

Linux (Ubuntu, Debian, Arch, Fedora)
macOS (Intel and Apple Silicon)
WSL2 on Windows
Android via Termux

Post-install step

Reload your shell to pick up the new hermes command:

source ~/.bashrc    # or: source ~/.zshrc

Verify installation

hermes --version

Expected output: Hermes Agent v1.x.x

Setup Wizard

Hermes includes an interactive setup wizard that guides you through initial configuration:

hermes setup

The wizard covers:

Model provider selection — Choose where your LLM comes from
API key entry — Securely store your credentials
Basic preferences — Set your default behavior

Model provider options

During setup, you’ll choose a model provider:

Provider	Models Available	Free Tier
Nous Portal	Hermes family	Yes (limited)
OpenRouter	200+ models	Yes
OpenAI	GPT-4o, GPT-4	No
Anthropic	Claude 4	No

Recommendation for first run: Select OpenRouter with a free model like openrouter:auto to test without cost.

First Interaction

Once setup completes, start Hermes:

hermes

You’ll see the TUI (Terminal User Interface) launch with a prompt.

Your first message

Type any message and press Enter. Hermes will respond using your configured model.

Example interaction:

You: Hello, what can you help me with?

Hermes: Hello! I'm Hermes Agent, your self-improving AI assistant. I can help you with:
- Answering questions and providing information
- Writing code and debugging
- Creating and managing tasks
- Learning your preferences over time

What would you like to explore?

Quick commands

Try these slash commands:

Command	Purpose
`/help`	Show available commands
`/model`	View/change current model
`/usage`	Check token usage
`/exit`	Exit Hermes

Troubleshooting

Issue 1: `hermes: command not found`

Cause: Shell not reloaded after install.

Fix:

source ~/.bashrc
# or restart your terminal

Issue 2: API key error

Cause: Missing or invalid API key for selected provider.

Fix: Run hermes setup again and enter a valid key.

Issue 3: Model not responding

Cause: Rate limit or network issue.

Fix:

Check /usage for rate limits
Try a different model with /model openrouter:auto
Verify network connectivity

Summary

You’ve completed the Hermes Agent quick start:

Installed via curl script
Configured via setup wizard
Verified with first interaction

Key takeaways

✅ Hermes installs in one command
✅ Setup wizard handles configuration
✅ First response confirms working installation
✅ Slash commands provide quick control

What’s next

Continue to Tutorial 2: Memory System to learn about Hermes’s unique persistent memory and user modeling.

Series navigation:

← Previous: Series Intro: Self-Improving AI Agent with Learning Loop
→ Next: Tutorial 2: Memory System — Persistent Memory and User Modeling
Back: Series Index