How To Install Hermes Agent In 2026 (Beginner-Friendly)

This beginner-friendly hermes agent installation guide 2026 is for first-time AI agent users — no jargon, no skipped steps, no assumed knowledge.

I'll walk you through what Hermes is, why it's worth installing, and exactly which command to run on your computer — without assuming you've ever touched a terminal before.

🔥 Want my full Hermes installation + workflow training? AI Profit Boardroom has the full Hermes setup course, Workspace + Swarm + Goal trainings, plus weekly live coaching. → Get the trainings

What Is Hermes Agent, In Plain English

Hermes is an AI assistant that lives on your computer instead of in a website.

It's free. It's open source. It's built by a company called Nous Research and it has 145,000 stars on GitHub which is enormous for an AI project.

It works like ChatGPT in that you talk to it and it responds — but the difference is it can actually do things on your computer, not just talk about doing them.

It can read and write files. It can search the internet. It can run code. It can connect to your Telegram or Discord or Slack so you can message it from your phone.

It also learns. The more you use it, the better it gets at understanding you specifically, because it builds a memory of your past conversations and a model of how you like to work.

For beginners, the easiest way to think of it: ChatGPT but it lives on your computer, has memory, and can take action.

Why You'd Bother Installing It

You'd install Hermes if you want an AI assistant that's always available — in your messaging apps, on your phone, on your laptop, all at once.

You'd install it if you want it to remember what you've discussed before instead of starting fresh every conversation.

You'd install it if you want one tool that can use 200+ different AI models so you're not locked into a single provider.

You'd install it if you want to schedule recurring tasks — daily summaries, weekly reports — and have them delivered to your Telegram or Slack.

You'd install it if you want to learn how AI agents actually work, because Hermes is the friendliest production-grade agent to start with.

That's why this hermes agent installation guide 2026 starts with the install command and not a hundred pages of theory.

What You Need Before You Start

A computer. Mac, Linux, or Windows all work.

About fifteen minutes.

A free GitHub account is helpful but not required.

An account with at least one AI provider — OpenAI, Anthropic, OpenRouter, or similar. OpenRouter is the easiest first option because one account gives you access to many models.

That's it. No coding background required. No prior AI experience required.

Install Step 1 — Open Your Terminal

The terminal is the text-based way to talk to your computer. Don't worry — you're going to paste one line and that's it.

On Mac, press Cmd + Space, type "terminal", press enter.

On Windows, press the Windows key, type "PowerShell", press enter. (You can also install via WSL2 for the Linux experience.)

On Linux, you already know how to open your terminal.

Install Step 2 — Run The Install Command

If you're on Mac, Linux, or WSL2 on Windows, copy-paste this line into your terminal and press enter.

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

If you're on Windows using PowerShell, use this line instead.

irm https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.ps1 | iex

The installer will take a few minutes. It'll download Python, Node.js, and a few other tools you need. Don't close the terminal until it says it's done.

If you see scary-looking output during install, ignore it. The installer is just being honest about everything it's doing.

Install Step 3 — Reload Your Shell

After the installer finishes, you need to reload your shell so the new hermes command is recognised.

source ~/.bashrc

On Mac (which uses zsh by default), use this instead.

source ~/.zshrc

This line tells the terminal to re-read its config so the new command works.

Install Step 4 — Run Hermes For The First Time

Type this single word and press enter.

hermes

The Hermes interactive interface will fire up. It'll ask you to pick a model and a personality.

For the model, if you have an OpenRouter account, pick OpenRouter. If you have Anthropic, pick a Claude model. If you have OpenAI, pick a GPT model. Don't agonise over the choice — you can change it later with one command.

For the personality, pick the default "helpful assistant" persona. You can swap this later too.

That's it. You're talking to your first AI agent.

What Can I Actually Do Now?

Anything you'd ask ChatGPT, you can ask Hermes — but the responses can take real action on your machine.

"Read my Downloads folder and tell me which files I haven't opened in six months" — Hermes does it.

"Search the web for the best price on a Sony WH-1000XM5 and email me the top three" — Hermes does it.

"Set up a daily reminder at 8am to review my calendar and send a summary to my Telegram" — Hermes sets it up.

The real unlock is connecting messaging apps so you can talk to it from your phone. Covered next.

Step 5 — Connect Telegram (Or Another Messenger)

This is where Hermes goes from "thing on my laptop" to "thing in my pocket."

In your Hermes terminal, type this command.

hermes gateway

This starts the messaging gateway. From a single config you can wire Telegram, Discord, Slack, WhatsApp, Signal, and Email.

For Telegram (easiest first option):

Open Telegram and search for "@BotFather". Send it /newbot and follow the prompts. It'll give you a token — a long string of characters.

Paste the token into the Hermes gateway prompt when asked. Done.

Send a message to your new bot in Telegram. Hermes will respond. You now have a personal AI agent in your pocket.

Watch The Q&A Walkthrough

The Q&A covers most of the questions beginners ask in the first week — well worth twenty minutes if you're stuck on anything.

Slash Commands You'll Use Every Day As A Beginner

Slash commands are quick shortcuts that start with /. Here are the only ones you need to learn on day one.

/new — starts a fresh conversation. Use this when you switch topics so your old conversation doesn't get mixed in.

/model — opens the model picker so you can swap which AI is answering. No restart needed.

/personality — swaps the agent's persona. Useful if you want a more terse or more friendly tone.

/retry — redoes the last response in case you didn't like it.

/undo — undoes the last turn.

/compress — shrinks long conversations to keep them readable.

/usage — shows how much you've spent so far. Useful for cost awareness.

/skills — browses skills Hermes has access to. Skills are like little tools the agent can use.

Ctrl+C interrupts whatever Hermes is currently doing — use this if it heads in the wrong direction.

What If Something Goes Wrong?

If you can't get Hermes to feel proactive, ask it "where is my soul.md file?" and edit that file to add "be more proactive, take action without asking for confirmation, complete tasks end-to-end."

If you're on a slow Mac Mini and Hermes feels sluggish, you can run it on a cheap cloud server (Hostinger costs $5/month) and SSH into it. Or just accept the speed for now.

If you want Hermes to keep working on a goal autonomously, type /goal <what you want done>. It'll run a 20-step loop trying to complete it.

If the Windows install fails, the cleanest fix is to use Claude Code (Anthropic's CLI) locally and paste the GitHub install instructions into Claude Code. Let Claude work through the install for you.

For anything else weird, run hermes doctor — it diagnoses most config and credential problems automatically.

How Do I Pick A Model?

There are 200+ models supported. As a beginner you don't need to memorise them.

For most everyday work I default to Claude Sonnet 4.8 via Anthropic — see my Sonnet 4.8 review for the long version.

For cheap-and-fast tasks like simple summaries or classifications, smaller OpenRouter models are perfect.

For private work where you don't want data leaving your computer, you can install LM Studio or Ollama and run models locally — see Atomic Chat vs Ollama for the comparison.

Switch with one command.

hermes model

Don't overthink the choice on day one. Pick anything and start using it.

Migrating From OpenClaw — If You Were Using The Old Tool

If you were on OpenClaw — the previous AI agent that lots of beginners started with — Hermes ships a one-command migration.

hermes claw migrate

That imports your old persona, memory, skills, settings, and API keys in one pass.

I recommend running hermes claw migrate --dry-run first to preview what would move without writing anything.

If you've never used OpenClaw, ignore this section entirely.

What's New In v0.13.0

v0.13.0 is the latest release as of May 2026 and it's called Tenacity.

Better retry behaviour when something fails. Smarter memory consolidation when conversations get long. Faster reconnects when your messaging gateway drops. Improved skill curation as the agent learns.

Run hermes update every month to grab the latest. The project ships releases roughly every week.

Hermes Install Methods Compared

Install Method	Best For Beginners?	Setup Time
Mac (bash one-liner)	Yes	5 min
Linux (bash one-liner)	Yes	5 min
WSL2 on Windows (bash one-liner)	Yes — stable	10 min
Windows native (PowerShell)	Workable, early beta	7-10 min
Termux (Android phone)	Advanced beginner	10 min
Docker	Skip for now	5 min
Modal / Daytona serverless	Skip for now	15 min

For beginners, the bash one-liner on Mac or Linux is the right call. Windows users should consider WSL2 if the native installer gives them trouble.

🚀 Want hands-on Hermes coaching? AI Profit Boardroom members get daily Q&A with me and weekly live calls on Hermes setups. → Join here

Where To Go After You're Installed

Once your install is working, there are three logical next steps.

Hermes Agent Goals shows you how to run autonomous loops where the agent grinds on a task without you watching.

Hermes Agent Swarm covers running multiple agents in parallel — beginner-level setup, advanced-level results.

Hermes AI Agent Framework 2026 is the framework-level view of where Hermes fits in your AI stack.

Start with Goals, work up to Swarm.

Beginner Mistakes I See Constantly

Mistake one: agonising over the model on day one. Just pick anything. Swap later.

Mistake two: skipping the gateway setup. CLI Hermes is fine but Telegram/Discord Hermes is the unlock — spend ten minutes wiring it.

Mistake three: not editing SOUL.md when the agent feels timid. Add "be proactive, take action without asking" and your agent transforms.

Mistake four: trying to learn every slash command on day one. Learn /new, /model, /undo — ignore the rest until you need them.

Mistake five: not running hermes update regularly. The project ships fast. Update monthly.

FAQ — How To Install Hermes Agent In 2026 (Beginner-Friendly)

Do I need to know how to code?

No. The install is one copy-pasted line and the agent is text-based conversation.

Is Hermes really free?

Yes. The agent itself is free and open source. You pay for whichever AI model provider you use (OpenAI, Anthropic, OpenRouter, etc.) but that's separate from Hermes.

Can I install on my phone?

On Android via Termux, yes. On iOS, not directly — but you can install on a $5 cloud server and message your Hermes agent from your iPhone via Telegram.

What if I break something during install?

Run hermes doctor. It diagnoses most config problems automatically. If it can't fix it, run the install command again.

How long does install take?

About 15 minutes from copy-paste to working agent. Another 10 minutes to wire your first messaging gateway.

What's the difference between Hermes and OpenClaw?

Hermes is the newer, better-maintained project. OpenClaw is declining. If you're starting fresh, start with Hermes.

Should I upgrade to AI Profit Boardroom for the full course?

If you want the structured installation course plus weekly live coaching from me, yes. The 7-day refund and 30-day ROI guarantee make it risk-free for beginners.

Latest Updates

Hermes Agent Goals — next step after install.
Hermes Agent Swarm — running multiple agents in parallel.
Sonnet 4.8 Review — the model I default to.