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

# Hermes Agent Template on Vertra Cloud: AI Agent with a Panel and Channels

> What the Hermes Agent template creates on Vertra Cloud: memory, what it asks, channels, variables and license. The full walkthrough is in the guide.

An AI agent that learns from use, with a web panel, an optional browser chat and messaging channels. The template runs Hermes Agent v0.22.5 behind a single login. You bring the model key: Vertra Cloud does not host a model.

> [card] [Step-by-step guide: Hermes Agent from the key to your channels](https://docs.vertracloud.app/templates/guides/hermes-agent.md)
>
> Provider key, Telegram, Discord, Slack and WhatsApp, the panel login, browser chat and changing the model.

## What it creates

One app (image). **Memory: 1.5 GB minimum**, 2 GB recommended. It needs a plan with Docker images and one free [image slot](https://docs.vertracloud.app/templates/overview.md#image-slots); see [Plans and limits](https://docs.vertracloud.app/plans-and-limits.md).

## What the wizard asks

| Question | What it does |
| - | - |
| Model provider | OpenRouter (default), Anthropic, OpenAI, Google Gemini or an OpenAI-compatible endpoint. |
| Provider API key | Kept only in the app's encrypted variables. |
| Base URL and model | The URL only with a compatible endpoint (ends in `/v1`). Empty model means the provider's default; required with a compatible endpoint. |
| Browser chat (Hermes WebUI) | Optional; about 150 MB of memory more. |
| Telegram, Discord, Slack, WhatsApp, email | Each asks for credentials and the **list of people allowed to talk to the agent**. |

## What you get

The panel at `<address>/admin/`, the user `admin` and a generated **password shown once**, and the next step. WhatsApp is paired by QR code in the panel.

## Variables

| Variable | What it does |
| - | - |
| `HERMES_ADMIN_USER`, `HERMES_ADMIN_PASSWORD` | The single login (`admin` and the generated password). |
| `HERMES_WEBUI` | `true` or `false`: the browser chat. |
| `HERMES_MODEL_PROVIDER`, `HERMES_MODEL`, `HERMES_MODEL_CUSTOM`, `HERMES_BASE_URL` | Provider and model, applied on every start. |
| `OPENROUTER_API_KEY`, `ANTHROPIC_API_KEY`, `OPENAI_API_KEY`, `GEMINI_API_KEY`, `OPENAI_COMPAT_API_KEY` | The chosen provider's key. |
| `TELEGRAM_*`, `DISCORD_*`, `SLACK_*`, `WHATSAPP_*`, `EMAIL_*` | Each channel's credentials and allowed list. |

## Resources and limits

* Port 8080. First boot 15 to 30 seconds; later starts about 15 seconds.
* State in `/app/.hermes`. The image is large, so the first deploy takes longer.
* Zero-downtime deploy is off; rollback returns the state too.
* An invalid channel token shows in the app log: do not share it.

## License

Hermes Agent is distributed under the [MIT license](https://github.com/NousResearch/hermes-agent/blob/main/LICENSE). The optional chat is Hermes WebUI (MIT) and the login proxy is Caddy (Apache-2.0). This template is a Vertra Cloud packaging and is not affiliated with the project.

See also the [public page](https://vertracloud.app/templates/hermes-agent) and the [other official templates](https://docs.vertracloud.app/templates/catalog.md).


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