Integration · Hermes

Use Verboo Code in Hermes

Hermes accepts OpenAI-compatible endpoints. Configure Verboo Code through the wizard or directly in its configuration file.

Start for free Technical documentation

Unlimited tokens

Paid plans have no token cap. What exists is a per-minute rate limit, because heavier models need more GPU.

Fixed price per developer

You know what you pay at the end of the month. An agent running all night does not become a surprise invoice.

Works with what you have

An OpenAI-compatible endpoint: swap the base URL and the key, keep your workflow.

How to connect

1
Setup wizard
- The recommended way to add an endpoint is through the Hermes terminal wizard.
- Select Custom endpoint (self-hosted / VLLM / etc.).
- Enter https://code.verboo.ai/router/v1 as the base URL.
- Paste your Verboo API key when prompted.
- Choose an ID returned by GET /models.
- Use the returned context_window as context_length and enable vision only when vision is true.
Open the wizard
```
hermes model
```

Manual configuration

Alternatively, edit ~/.hermes/config.yaml and define the model, provider, base URL, and API key.

~/.hermes/config.yaml

model:
  default: <MODEL_ID>
  provider: custom
  base_url: https://code.verboo.ai/router/v1
  api_mode: chat_completions
  api_key: <VERBOO_API_KEY>
  context_length: <CONTEXT_WINDOW>
  supports_vision: <VISION_ENABLED>

3
Verify
- Run Hermes diagnostics after saving. It reports unknown providers, missing credentials, and configuration problems.
Run diagnostics
```
hermes doctor
```

Keep the key private

If your workflow uses shared files, inject the credential through Hermes secrets rather than saving it as plain text.

Model capabilities

Replace placeholders with data from the models page. supports_vision must be false for models without vision. When reasoning is available, use only effort levels returned by the Router.