Hermes Agent CLI Cheatsheet - Nous Terminal Multi-Backend AI Agent

All Hermes Agent terminal AI-agent commands in one place — from one-line install to model/tool config, gateway, and setup wizard — organized by scenario. Copy and go.

AI CLI·15 commands·Last updated 2026-08-26

Install & Launch 6

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

One-line installer for Linux/macOS/WSL2

pipx install hermes-agent

Install into an isolated environment via pipx

pip install hermes-agent

Install directly with pip

hermes

Start the interactive TUI chat

hermes --version

Print the installed version

hermes doctor

Diagnose missing dependencies and misconfigurations

Model & Tool Config 5

hermes model

Launch the model-backend wizard (OAuth / API key / custom endpoint)

hermes tools

Configure which built-in tools are enabled

hermes config set <key> <value>

Set a single configuration value

hermes config set terminal.backend docker

Switch the execution backend to a docker sandbox

hermes config set terminal.backend ssh

Delegate heavy work to a remote host

Gateway & Wizard 4

hermes setup

Full setup wizard (model + tools in one flow)

hermes gateway

Start the message gateway (Telegram / Discord, etc.)

hermes chat

Start a chat session directly

docker run --rm -it -v ~/.hermes:/root/.hermes -e OPENAI_API_KEY=$OPENAI_API_KEY nousresearch/hermes-agent:latest

Run an isolated instance via Docker

Tips

Hermes is model-agnostic: run `hermes setup` once to wire up 20+ backends (OpenRouter, Kimi, DeepSeek, OpenAI) instead of configuring each tool separately.
Before running tasks in untrusted environments, set `hermes config set terminal.backend docker` to isolate execution in a sandbox rather than touching the host.
Run `hermes doctor` first when something breaks — it lists missing dependencies and config issues directly, saving blind retries.

FAQ

Which model backends does Hermes Agent support?

It supports 20+ backends including OpenRouter (100+ models), DeepSeek, Kimi/Moonshot, OpenAI, Google Gemini, Groq, Ollama, vLLM, and any OpenAI-compatible endpoint, configured via the `hermes model` wizard on first run.

How does Hermes Agent isolate its execution environment?

Switch the execution backend with `hermes config set terminal.backend`: local runs on the host, docker sandboxes tool execution, ssh delegates to a remote host, and daytona/modal provide ephemeral dev environments or serverless GPU.

Official References

Each command links to its official documentation below, so you can verify the latest usage and read deeper.

Maintained by LaoHand

Publicly updated on Aug 26, 2026, continuously proofread against official docs.

Contact Us

Wrong command or description? Send us corrections, business inquiries or product feedback by email.

Edit this page on GitHub

Edit the source file directly and open a PR to improve this cheatsheet.

Edit this page on GitHub Report an issue