How to Install Hermes Agent in 60 Seconds
Step-by-step guide to installing Hermes Agent on Linux, macOS, or VPS with the official installer.
Installing Hermes Agent takes under a minute — a single curl command handles Python, dependencies, and the CLI automatically. Once installed, you are talking to your agent from the terminal or any connected channel.
Before you start:
- ☑Linux, macOS, or WSL2 (Windows Subsystem for Linux)
- ☑Git installed (required by the installer)
- ☑An API key from Nous Portal, OpenRouter, OpenAI, or Anthropic — OR Ollama for fully local inference
Steps
- 1
Run the installer
curl -fsSL https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.sh | bash
- 2
Configure your model
Edit config.yaml and add your API key (Nous Portal, OpenRouter, or OpenAI)
- 3
Start Hermes
hermes start
Pro Tips
- 💡After install, run 'source ~/.bashrc' (or ~/.zshrc) before typing 'hermes' — otherwise the command won't be found
- 💡Run 'hermes model' after first install to pick your LLM provider interactively instead of editing config files manually
- 💡Start Hermes from your home directory (~/) to minimize token overhead from workspace context injection
Troubleshooting
❌ Command 'hermes' not found after install
✅ Run 'source ~/.bashrc' or 'source ~/.zshrc' to reload your shell PATH. The installer adds Hermes to PATH but doesn't reload the current shell automatically.
❌ Installer fails with 'git not found'
✅ Install git first: 'sudo apt install git' on Debian/Ubuntu or 'brew install git' on macOS, then re-run the installer.
❌ API key rejected on first run
✅ Check your API key for URL-encoding artifacts (e.g. %3D instead of =). Copy the raw key from the provider dashboard and paste it directly into config.yaml.