Singularity.Kiwi AI progress, with honesty.
← Back to Toolkit

⚡ Set Up Hermes Agent

The fastest path to a working AI agent. 15 minutes, beginner-friendly.

🕐 Time 15–30 min
📊 Difficulty Beginner
💰 Cost $0–20/mo
💻 Platform Mac · Win · Linux

What is Hermes Agent?

Hermes Agent is an open-source AI agent by Nous Research. It runs on your computer, connects to your chat apps, and has persistent memory — it remembers you across sessions.

🧠 Persistent memory Remembers you, your preferences, and your environment across sessions.
🔧 Native tools File access, web browsing, terminal — built in, no plugins needed.
💬 Multi-platform Telegram, Discord, built-in chat. Use it wherever you already talk.
📝 Self-improving Learns from experience by saving reusable skills. Gets better over time.
⏰ Scheduling Built-in cron. Run tasks on a schedule, fully automated.
🔄 Provider-agnostic Swap models and providers mid-workflow. Ollama, OpenAI, Anthropic, 20+ more.

Visit hermes-agent.nousresearch.com →

Step 1: Install Ollama

Hermes needs an AI model to think with. The easiest way is Ollama Cloud — sign up free, grab an API key, and paste it during setup.

1

Download the Ollama app

Go to ollama.com and download the app for your operating system. Install it — takes about a minute. Run the app once and sign in.

Download Ollama →
2

Get your API key

Sign up for a free account at ollama.com → Settings → API Keys → Create API Key. Copy the key and save it somewhere safe — Ollama only shows it once.

3

Pick a plan

Free $0 Testing & light use. Hits limits quickly during longer tasks.
Pro $20/mo 50x more usage. Larger models. Run 3 at once. Sweet spot for daily use.
Max $100/mo 5x Pro. 10 models at once. Only for heavy sustained agent work.

💡 Our recommendation: Start on Free to verify the install, then upgrade to Pro. At $20/month, Pro is enough for most users to run a single agent daily without mid-task cutouts.

Don't want Ollama? You can also use OpenAI (GPT-4o, pay-as-you-go, ~$2.50–10/1M tokens) or Anthropic (Claude, pay-as-you-go). You only need one provider. Ollama Cloud is the cheapest way to start.

Step 2: Install Hermes

Hermes handles Python and dependencies automatically. You don't need to install anything else first.

A

Mac Desktop App Easiest

  1. Make sure the Ollama app is installed and running
  2. Go to hermes-agent.nousresearch.com
  3. Download the Mac app
  4. Drag it to your Applications folder
  5. Open it
  6. Follow the setup wizard:
    • Choose your AI provider (Ollama Cloud, OpenAI, or Anthropic)
    • Paste your API key
    • Pick a model
    • Choose whether to enable Telegram, Discord, or just use the built-in chat
  7. Start chatting

No terminal required.

B

Terminal Command Windows · Linux · Mac

  1. Make sure the Ollama app is installed and running
  2. Open your terminal
  3. Paste this command and press Enter:
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash

The installer sets up Hermes, its Gateway, and dependencies automatically.

When it finishes, run the setup wizard:

hermes setup

The wizard asks you to:

  • Choose chat platform (Telegram recommended, or built-in chat)
  • Choose AI provider
  • Paste your API key
  • Pick a model

Then start the Gateway:

hermes gateway start

Step 3: Test It

✅
Gateway running?
hermes gateway status

Should say "running."

✅
Chat working?

Open the Hermes app or message your bot and say "Hello".

✅
AI working?

Ask "What's 2+2?" — should say 4.

✅
Memory working?

Say "My name is Chris", then in a new chat ask "What do you know about me?" — it should remember.

Step 4: Switch Models (Once Running)

With the Ollama app running and your API key set up, the fastest way to try a different model:

ollama launch hermes --model glm-5.2:cloud

This downloads the model if you don't have it yet and reconnects Hermes in one step. Browse available models at ollama.com/library.

At publishing time, glm-5.2:cloud is excellent for publishing, research, and content work.

You can also switch models in the Hermes app settings, or by asking: "Switch to glm-5.2:cloud."

Step 5: Make It Start Automatically

Hermes can run as a background service so it survives reboots.

Desktop app: Look for "Start on login" or "Run as service" in settings.

Terminal:

hermes gateway install

Then check:

hermes gateway status

If that says "running," you're done.

Optional: Add Telegram

You don't need Telegram — the built-in chat works fine. But once your agent works, adding Telegram lets you chat from your phone.

Just ask your agent: "Connect me to Telegram. Walk me through it."

Or do it manually:

  1. Open Telegram, search @BotFather, send /newbot
  2. Pick a name and username (must end in "bot")
  3. Copy the token
  4. Run hermes gateway setup and paste the token

Hermes vs OpenClaw

Feature OpenClaw Hermes
Install One-liner script Desktop app or one command
Memory Session + file-based Built-in persistent memory
Tools Plugin system Native file, web, terminal
Scheduling Built-in cron Built-in cron
Chat apps Telegram, Discord, iMessage, etc. Telegram, Discord, built-in chat

💡 In plain English: OpenClaw is great if you want maximum chat-app coverage. Hermes is great if you want the cleanest setup and strongest memory out of the box.

Prefer OpenClaw? OpenClaw setup guide →

You're Ready

Download the Mac app, or run one terminal command. Answer the wizard. Paste your API key. Start chatting.

Hermes is the fastest path from zero to a working agent.

📖 Get The Lobster Farm (Free) → OpenClaw Setup → Back to Toolkit →

Want more help? The Lobster Farm is our free ebook that walks you through building your own local AI agent stack — the same approach we used to build Singularity.Kiwi. 12 chapters, practical, beginner-friendly.