Quick start

HermesAnsem is a self-contained agent — one command installs it as hermesansem on your machine. No accounts, no sign-up. From zero to your first conversation in about a minute.

Prerequisites

• OS — macOS, Linux, or WSL2 on Windows.
• Node.js 18+ — the agent runs on it (check with node -v).
• API key — OpenAI, Anthropic, OpenRouter, or any OpenAI-compatible endpoint.

1 · Install

Easiest — straight from npm, nothing to clone (needs Node.js 18+). Run it with no install at all:

npx hermesansem setup

…or install the hermesansem command globally, so you can drop the npx prefix everywhere below:

npm install -g hermesansem

No npm? The one-line script installs a standalone CLI — then reload your shell:

curl -fsSL https://hermesansem.xyz/install.sh | bash
source ~/.zshrc    # or: source ~/.bashrc

2 · Configure

Run the setup wizard — pick a provider and paste your key (saved to ~/.hermesansem/config.json):

hermesansem setup

3 · Run

Start the agent:

hermesansem

You're now talking to HermesAnsem in your terminal. In-chat commands: /new for a fresh conversation, /exit to quit. Your conversation is remembered in ~/.hermesansem.

FAQ

Do I need to write code? No. Install, run hermesansem setup, then hermesansem and start chatting.

Which model should I use? Any OpenAI-compatible model. Change it any time with hermesansem setup.

Where is my data? Config and memory live locally in ~/.hermesansem. Nothing leaves your machine except the calls to your chosen model.

Chat with the agent How it works