> ## Documentation Index
> Fetch the complete documentation index at: https://ngrok.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Hermes

> Install the ngrok AI Gateway plugin for Hermes Agent to use your gateway's models as a Hermes model provider with one access key.

[Hermes Agent](https://hermes-agent.nousresearch.com) is Nous Research's open-source AI agent for the terminal.
The [ngrok AI Gateway plugin](https://github.com/ngrok/hermes-ngrok) adds the gateway as a Hermes model provider, so Hermes can use every model your access key reaches.

## What you'll need

* [Hermes Agent](https://hermes-agent.nousresearch.com) 0.21.6 or later installed.
* An [access key](/docs/ai-gateway/concepts/access-keys) from [app.ngrok.ai](https://app.ngrok.ai/keys).

## Install the plugin

Install the plugin with the Hermes CLI:

```bash theme={null}
hermes plugins install ngrok-ai-gateway
```

Enter your access key when prompted.
Hermes saves it as `NGROK_AI_GATEWAY_KEY`.

Plugins are installed per Hermes profile, so install it again in any other profile that needs it.

## Choose a model

Run `hermes model` and select **ngrok AI Gateway**, or pass `--provider ngrok` when you start Hermes.

The model list comes from your gateway, so it shows the models your access key is allowed to call, by full catalog ID such as `openai/gpt-5.5`.
Model IDs follow the gateway's [model naming](/docs/ai-gateway/concepts/models): a model ID such as `gpt-5.5`, or `provider:model` to pin a provider, such as `openai:gpt-5.5`.

## Test the connection

Send a single prompt from the command line:

```bash theme={null}
hermes chat --provider ngrok --model gpt-5.5 -q "Explain what a hash map is in two sentences."
```

Confirm the request in AI Gateway [observability](/docs/ai-gateway/observability) in [app.ngrok.ai](https://app.ngrok.ai).

## Use a cheaper model for background tasks

Background tasks like context compression and title generation use your main model.
To send them to a cheaper model through the gateway, set the provider and model for each task in `~/.hermes/config.yaml`:

```yaml theme={null}
auxiliary:
  compression:
    provider: ngrok
    model: gpt-5.4-mini
  title_generation:
    provider: ngrok
    model: gpt-5.4-mini
```

## What the plugin sends

The plugin sends chat and model list requests to `https://gateway.ngrok.ai/v1` with your access key.
Each request carries a `User-Agent: hermes-agent/<version> (ngrok-plugin)` header.
The plugin makes no other network requests, doesn't read files, and doesn't run background processes.

## Next steps

* [Restrict providers and models](/docs/ai-gateway/guides/restrict-providers-and-models): Scope what an access key can call
* [Configure fallback models](/docs/ai-gateway/guides/configure-fallback-models): Try another model when the first one fails
* [Observability](/docs/ai-gateway/observability): See which model served each request


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.