Skip to main content
Hermes Agent supports custom endpoints that speak the OpenAI API. Use its interactive model setup to keep the Aixy key in Hermes’ local configuration.

​
Configure a custom endpoint

  1. Run hermes model.
  2. Select Custom endpoint.
  3. Enter https://api.aixy-gateway.com/v1 as the base URL.
  4. Enter a project-scoped Aixy API key when prompted.
  5. Choose Chat Completions when prompted for the API mode.
  6. Select an available model, such as openai/gpt-4.1-mini, or an aixy/<route>, and supply its verified context limit when prompted.
Hermes saves this provider in its local configuration. Use hermes model again to switch the active model without placing the key in a repository. Send a short prompt and verify the request in Activity. Review any auxiliary-model overrides in Hermes so you know which connection handles those tasks.
Do not set an Aixy key in a shared shell profile or commit it in config.yaml. Prefer the credential prompt or a local secret store managed by Hermes.
See the Hermes custom-endpoint documentation for provider-specific options.