Skip to content
VikeyVikey Docs

Hermes

View as Markdown

Hermes Agent calls Vikey through the OpenAI chat completions format.

Tested with Hermes Agent 0.21.2 on 12 November 2026.

Steps

Section titled “Steps”

  1. Install Hermes if you do not have it:

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

  2. Keep the API key in an environment variable:

    Terminal window
    export VIKEY_API_KEY=YOUR_API_KEY

  3. Add the Vikey provider and make it the default model in ~/.hermes/config.yaml:

    ~/.hermes/config.yaml
    model:
      default: vikey/agent-gpt
      provider: vikey
      base_url: https://api.vikey.ai/v1
      key_env: VIKEY_API_KEY
      api_mode: chat_completions
    

    providers:
      vikey:
        name: Vikey
        base_url: https://api.vikey.ai/v1
        key_env: VIKEY_API_KEY
        model: vikey/agent-gpt
        models:
          vikey/agent-gpt: {}
          openai/gpt-6.1-sol: {}
          deepseek/deepseek-v4.1-flash: {}

  4. Run hermes. To change models, run hermes model, choose Vikey, then the model.

Another way, without editing the file:

Terminal window
hermes config set providers.vikey.name "Vikey"
hermes config set providers.vikey.base_url "https://api.vikey.ai/v1"
hermes config set providers.vikey.key_env "VIKEY_API_KEY"
hermes model

Common problems

Section titled “Common problems”
What you see Cause Fix
Connection error or a timeout in Hermes Desktop (macOS) A TLS problem in Hermes Desktop’s built-in HTTP client, not in your config or the Vikey API Temporarily set base_url to http://api.vikey.ai/v1, restart Hermes Desktop, and switch back to https once Hermes fixes it
401 VIKEY_API_KEY is not read Export it again and start Hermes from the same terminal
Model '...' not found A typo in the model id Copy the id from the pricing page