Hermes Agent - Self-improving AI Agent

This guide explains how to install Hermes Agent and configure Lychee AI as its custom model provider.

Before you begin

Prepare the following information:

1. API URL: https://lizh.ai/v1
2. API key: Sign in to Lychee AI, then create and copy a key from Console → API Keys
3. Model ID: Copy the exact model name you want to use from the Lychee AI model marketplace

Install Hermes Agent

macOS, Linux, or WSL2

Run the official Hermes installer in a terminal:

curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash

Windows

Run the following command in PowerShell:

iex (irm https://hermes-agent.nousresearch.com/install.ps1)

After installation, open a new terminal window and check the installed version:

hermes version

If version information appears, Hermes has been installed successfully.

Configure Lychee AI

Hermes recommends using its model setup wizard to add a custom OpenAI-compatible endpoint.

1. Open the model setup wizard

Run this command in your terminal:

hermes model

2. Select a custom endpoint

From the Provider list, select:

Custom endpoint (self-hosted / VLLM / etc.)

The exact wording may vary by version. Choose the option containing Custom endpoint.

3. Enter the connection details

Follow the prompts and enter:

If the wizard asks for a Context length, enter the context length shown for the model. If you are unsure, keep the default value provided by Hermes.

When setup is complete, Hermes saves the model, provider, and API URL to ~/.hermes/config.yaml, so the configuration remains available after restarting your terminal.

Verify the connection

Run:

hermes chat

Send a simple message after the chat starts. If Hermes returns a response, the API URL, API key, and model ID are configured correctly.

To change the Lychee AI model later, exit the current session and run:

hermes model

Manual configuration (optional)

If you cannot use the interactive wizard, edit ~/.hermes/config.yaml:

model:
  default: "model ID copied from the model marketplace"
  provider: custom
  base_url: https://lizh.ai/v1
  api_key: "your lizh.ai API key"

Save the file and run hermes chat again. YAML requires spaces for indentation, not tabs. Never commit a configuration file containing a real API key to a public repository.

Troubleshooting

401 or Authentication failed

The API key is invalid, expired, or incomplete. Check the key status in Lychee AI API Key Management and create a new key if necessary.

404, Model not found, or model unavailable

The model ID must exactly match the name shown in the Lychee AI model marketplace. Check capitalization, hyphens, and version numbers.

Cannot connect to the API

Confirm that the API Base URL is https://lizh.ai/v1, including /v1, and verify that your network can reach Lychee AI.

Lychee AI does not appear under /model

Exit the Hermes chat session and run hermes model in your system terminal to configure the custom endpoint first. After setup, you can use /model inside a session to switch models.

hermes: command not found

Close and reopen your terminal. If the command is still unavailable, follow the Hermes documentation to check the installation path and PATH configuration: FAQ & Troubleshooting.