Hermes Guide

Hermes is an AI Agent tool that supports connecting to TokenFlux through a custom OpenAI-compatible interface.

Installation

Choose the official installer for your system:

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

iex (irm https://raw.githubusercontent.com/NousResearch/hermes-agent/main/scripts/install.ps1)

Open a new terminal after installation so PATH changes take effect. The default configuration directory is:

System	Configuration directory
macOS / Linux / WSL	~/.hermes
Native Windows installation	%LOCALAPPDATA%\hermes

If HERMES_HOME is set, use the directory it specifies. Both config.yaml and .env below are in this configuration directory.

Connect to TokenFlux

1. Follow Create API Key to generate an API key.

2. Edit config.yaml in the configuration directory and add or confirm the following:

   model:
     default: 'gpt-6-astra'
     provider: 'custom'
     base_url: 'https://tokenflux.dev/v1'

3. Edit .env in the same directory and write:

   OPENAI_API_KEY=YOUR_TOKENFLUX_API_KEY
   OPENAI_BASE_URL=https://tokenflux.dev/v1

   Replace YOUR_TOKENFLUX_API_KEY with your actual API key.

Verify Configuration

After saving the configuration, run:

hermes config check
hermes chat -Q -q 'Only reply OK' --max-turns 3

hermes config check checks the local configuration. hermes chat makes a real, billed model request. After receiving a reply, confirm the request in the usage logs.

If the client says the API key was rejected, read the server message after Provider said. Billing errors, such as an unavailable subscription bound to the key, can appear under that heading too. See Error Codes and Balance and Billing for the corresponding action.

More Related Content

• Create API Key
• Billing