Skip to main content
Hermes is an agent tool from Nous Research. It can connect to a custom OpenAI-compatible endpoint, so you can use Novita LLM models inside Hermes for local agent workflows.

​
Prerequisites

Before you begin, prepare the following values:
  • API key: Create or copy an API key from Key Management.
  • Base URL: https://api.novita.ai/openai/v1
  • Model ID: Copy the model ID you want to use from the Model Hub, such as deepseek/deepseek-v3.1.
Hermes stores endpoint and API key settings locally. Do not commit any generated configuration that contains your API key.

​
Install Hermes

Hermes provides terminal and desktop installation options. Choose the one that fits your workflow.

​
Linux / macOS / WSL2

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

​
Windows PowerShell

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

​
Desktop

You can also download Hermes Desktop from: 
https://hermes-agent.nousresearch.com/desktop

​
Configure the Novita endpoint

After installation, run the Hermes setup flow. The exact command may vary by Hermes version, so follow the terminal prompts from your installed version. During setup, use the following choices and values:
  1. Select Custom endpoint or the option that lets you enter a provider URL manually.
  2. Set the Base URL to https://api.novita.ai/openai/v1.
  3. Enter your Novita API key.
  4. For API type, keep the default auto-detection option. If you need to choose manually, select the OpenAI-compatible or Chat Completions mode.
  5. Select a model from the list, or enter a Novita model ID manually, such as deepseek/deepseek-v3.1.
  6. Set max tokens based on the model context length. If you are not sure, keep the default or auto-detected value.
  7. Give the model configuration a recognizable name, such as novita-deepseek-v3.1.
  8. If you only want to run Hermes locally, choose the default Local runtime.
  9. If you do not need channels such as Telegram, Slack, or Matrix, skip integration channels.

​
Reload your shell

If Hermes asks you to reload your shell, run: 
source ~/.zshrc

​
Start Hermes

Start Hermes from your terminal: 
hermes
Send a test message, for example: 
Hello, briefly describe what coding tasks you can help with.
If the request fails, check that:
  • Your API key is correct.
  • The Base URL is https://api.novita.ai/openai/v1.
  • The model ID matches the one shown in Novita’s Model Hub.
  • The selected API type is compatible with an OpenAI-compatible endpoint.
Last modified on June 25, 2026