Deploy and access an AI agent · Galaxium Hosting help
Deploy OpenClaw, Hermes, Paperclip, OpenCode or OpenHands pre-wired to your AI Router credits, then sign in and manage it.
Galaxium Hosting runs ready-made AI agents for you. Each one can be connected to the AI Router automatically, so it uses your AI credits and you don't need a separate provider account.
| Agent | What it is | How you sign in | |---|---|---| | OpenClaw | A personal AI agent on WhatsApp, Telegram, Discord, Slack and more | Web Control UI with a gateway token | | Hermes Agent | A self-improving agent with memory, chat apps and scheduled tasks | SSH (no web interface) | | Paperclip | Run a company of AI agents with an org chart, goals and budgets | Web app with an owner account and password | | OpenCode | An open-source coding agent in your browser, with a persistent workspace | Browser login: user opencode and a password | | OpenHands | Agent Canvas: coding conversations and scheduled automations | Web app that asks for a session API key |
Some agents run on a VPS of your own instead (they appear under On your own VPS): goose and Codex CLI are terminal agents you use over SSH (Codex signs in with your own OpenAI account and can't use the AI Router), and ZeroClaw, AdClaw and Haystack (Hayhooks) have dashboards without a full login, so they listen on localhost and you open them through an SSH tunnel.
Deploy
- Open AI agents in the sidebar and choose the agent.
- Click Deploy (the button shows the agent's name).
- Enter an Instance name and choose the Region and Default model.
- Under AI connection, keep Galaxium AI Router (recommended), or choose Bring your own provider key.
- Click Deploy for $X/mo.
- The admin login (token or password) is shown once. Save it, then continue to your agent.
With the AI Router option, a dedicated API key is created for the agent and set as its OPENAI_BASE_URL and OPENAI_API_KEY. The key also appears under AI Router → API keys.
Access
- Overview has an Open … ↗ button for agents with a web interface.
- Access & security shows the SSH login for Hermes, lets you reset the token or password (the new one is shown once), add a Custom domain with a CNAME record and, for web agents, add HTTP basic auth and an IP allow-list.
Manage
- AI & models: change the default model, set a Monthly credit cap for the agent's key, restrict models or rotate the key. A model change or key rotation needs a restart; caps and restrictions apply right away.
- Channels (OpenClaw, Hermes) connect chat apps. Companies & agents (Paperclip) manages your AI company. Integrations (OpenCode, OpenHands) explains where tokens for GitHub and other tools go.
- Environment: add your own variables (
OPENAI_*names are reserved). - Backups: daily backups plus manual ones. A safety backup is taken before updates, restores and reinstalls.
- Logs: live tail and download.
- Settings: version and automatic updates, reinstall, and delete.
What about n8n?
n8n runs on a VPS that you control. On its page, click Deploy n8n to order a VPS with n8n installed (Docker on Ubuntu 24.04, n8n on port 5678). Galaxium Hosting supports the VPS; you manage n8n inside it. To use your AI credits in n8n, add an OpenAI credential with your AI Router base URL and a gx- key.