Now live: DeepSeek V4.1 Flash and GLM-5.3, hosted in Australia. Get started →

Integrations Hermes Agent

Use DeepSeek in Hermes, hosted in Australia

Point Hermes Agent at the Xinference Model API and run DeepSeek V4.1 Flash or GLM-5.3 in Australian data centres. Four steps, no code.

Hosted in Australia

Set it up in four steps

  1. 1

    Get your key

    Sign up for the Model API and create an API key in the console.

  2. 2

    Open the model picker

    In a terminal, outside any chat session, run this and choose Custom endpoint (self-hosted / VLLM / etc.).

    Terminal
    $ hermes model
  3. 3

    Enter your details

    Paste these three values when Hermes asks. Leave context length blank. If Hermes asks for the API mode, choose chat_completions.

    Hermes custom endpoint
    Base URL  https://cloud.xinference.co/v1
    API key   your Xinference API key
    Model     deepseek-v4.1-flash
  4. 4

    Start a session

    Hermes now runs on DeepSeek, hosted in Australia. If anything fails, hermes doctor checks your setup.

    Terminal
    $ hermes chat

Prefer a config file?

Put the same three values in ~/.hermes/config.yaml.

~/.hermes/config.yaml
model:
  default: deepseek-v4.1-flash
  provider: custom
  base_url: https://cloud.xinference.co/v1
  api_key: XINFERENCE_API_KEY

To switch models in a session, type /model custom:glm-5.3.

Run Hermes on DeepSeek, hosted in Australia