Configuring Hermes

Hermes is a command-line agent from Nous Research that connects to models over the OpenAI-compatible protocol.

1. Install

For Linux / macOS / WSL2 (Git is the only prerequisite; the install script handles the rest):

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

After installation, reload your shell (or open a new terminal).

2. Run the setup wizard

hermes setup

Follow the prompts:

1. Choose Quick Setup;
2. Model provider: choose the Custom / OpenAI-compatible option;
3. Base URL: enter https://api.wylon.cn/v1;
4. API Key: paste your wylon API Key;
5. Model name: enter the model ID you want, e.g. moonshotai/Kimi-K2.6;
6. Complete the remaining options.

3. Use

Once configured, run the command Hermes shows you to start (for example, hermes to enter an interactive session).