Hermes Agent Installation & Usage
This chapter walks you through installation, initial configuration, and your first conversation.
System Requirements
- OS: Linux, macOS, WSL2, or Termux on Android
- Network: access to your chosen LLM provider (Nous Portal / OpenRouter, etc.)
- The installer automatically sets up uv, Python 3.11, Node.js, ripgrep, ffmpeg, and bundled Git (MinGit)
You do not need to install Python manually — the one-line script prepares the runtime for you.
One-Line Install
Linux / macOS / WSL2 / Termux:
Windows (PowerShell):
The script clones the repo, creates a virtual environment, installs dependencies, and registers the hermes command.
From Source (developers)
Manual equivalent:
Initial Configuration: hermes setup
After installing, run the setup wizard:
It guides you through:
- Choosing a provider (Nous Portal / OpenRouter / local vLLM / custom endpoint)
- Entering credentials (OAuth login or API key)
- Picking a default model
- Selecting tools to enable
- (optional) Configuring the message gateway
To do an individual step, use the focused commands:
Your First Conversation
Run hermes to enter the interactive terminal UI:
You'll see a chat interface. Try:
Then give it a real task:
Hermes will request the relevant tools (write file, run command) and execute them after you approve.
Common CLI Commands
Update & Diagnose
When something breaks, run hermes doctor first — it inspects your runtime, dependencies, and config and suggests fixes.
Migrating from OpenClaw
If you previously used OpenClaw, migrate personas, memories, skills, and keys in one step:
Next Steps
- Model Configuration — connect and switch LLMs
- Slash Commands — in-chat shortcuts
- Hands-on Practice — build a personal assistant from scratch