Managed Hermes hosting 

# Run Hermes. Skip the server work.

Get a dedicated Hermes agent with persistent state, a protected web dashboard, Telegram, and managed model routing—without maintaining the infrastructure yourself.

[Deploy Hermes](/users/sign_up)[See how it works](#how-it-works)

Choose Hermes during setup · Uses existing plans · Cancel anytime

HHermes dashboard
  Running 

 Managed runtime 

Telegram

Memory & skills

AI models

## Keep Hermes. Hand off the infrastructure.

Self-hosting gives you control, but it also gives you every operational task. Managed hosting keeps the Hermes experience while we run the deployment around it.

1
### When you self-host

- Provision and secure a server
- Configure the runtime and public access
- Keep state and skills on persistent storage
- Diagnose process and deployment failures

2
### With OpenClaw Direct

- A dedicated managed Hermes runtime
- Persistent configuration, memory, and skills
- Authenticated access to the web dashboard
- One portal for setup, billing, and instance controls

## What managed Hermes hosting includes

The supported surfaces you need to start using Hermes without assembling the hosting stack yourself.

### Dedicated Hermes runtime

Your instance runs Hermes in its own managed deployment rather than on a laptop or shared browser session.

### Persistent memory and skills

Hermes configuration, memory, and installed skills live on persistent storage across normal runtime restarts.

### Protected web dashboard

Open the Hermes dashboard through the existing authenticated portal path instead of exposing it directly to the internet.

### Telegram support

Connect a Telegram bot during setup and use Hermes from a chat you already keep open.

### Managed model routing

Choose from models supported by the platform AI gateway and configure provider access through the setup flow.

### ChatGPT and Codex pairing

Active Hermes instances can start the supported device-pairing flow for ChatGPT or OpenAI Codex access.

## From signup to Hermes in three steps

Use the same guided flow as every managed instance—then choose Hermes as the runtime.

1

### Choose a plan

Create your account and select the plan that fits the resources you need.

2

### Select Hermes

Pick Hermes, choose supported model access, and optionally connect Telegram.

3

### Open your agent

Launch the protected web dashboard or start a conversation from Telegram.

## A small, managed path to your agent

Requests reach your dedicated Hermes runtime through managed access and model-routing layers, while state remains on persistent storage.

You

Telegram

H

Hermes runtime

Managed runtime

Persistent state

AI gateway

## Clear about what works today

Hermes and OpenClaw are different runtimes. We only advertise integrations wired for the runtime you choose.

Hermes is developed by Nous Research. OpenClaw Direct provides the managed hosting service.

### Supported with Hermes

- ✓ Authenticated web dashboard 
- ✓ Telegram 
- ✓ Supported AI gateway models 
- ✓ ChatGPT and Codex device pairing 

### Need Slack?

Slack is not currently available with Hermes. Choose OpenClaw for a managed agent that connects to Slack.

[Explore OpenClaw for Slack](/slack)

### One set of plans, your choice of runtime

Hermes uses the existing managed-hosting plans. Compare resources and billing options on the pricing page; you choose Hermes during setup.

[View pricing](/pricing)

## Managed Hermes hosting FAQ

What is managed Hermes hosting?

It is a dedicated Hermes agent runtime operated through OpenClaw Direct. We manage the deployment and access path while you configure and use Hermes.

Do I need to run my own server?

No. You use the web portal, protected Hermes dashboard, and supported messaging channel without provisioning a VPS or keeping a computer awake.

What persists on my Hermes instance?

The managed runtime stores Hermes configuration, memory, and skills on persistent storage so they remain available across normal runtime restarts.

How do I access the Hermes dashboard?

Active Hermes instances expose the dashboard through the existing authenticated portal path. It is not presented as an unauthenticated public dashboard.

Can I use Hermes from Telegram?

Yes. You can configure a Telegram bot and allowed user during setup, then message your Hermes instance from Telegram.

Which AI models and API keys can I use?

Hermes can use models supported by the OpenClaw Direct AI gateway. The setup flow shows the currently available managed and provider-key options.

Does ChatGPT or Codex pairing work with Hermes?

Yes. For an active Hermes instance, the portal can start the supported device-pairing flow for ChatGPT or OpenAI Codex.

Does Hermes support Slack?

Not currently on OpenClaw Direct. If Slack is required, choose the OpenClaw runtime and use the dedicated Slack integration.

How does pricing and cancellation work?

Hermes uses the existing managed-hosting plans and normal billing flow. See the pricing page for current plan details; subscriptions can be cancelled through the existing account controls.

## Use Hermes without becoming its sysadmin

Create an account, choose Hermes during setup, and let the managed platform handle the deployment around it.

[Deploy Hermes](/users/sign_up)[Compare plans](/pricing)