Configure OpenClaw Remote Gateways and Nodes

Ever wanted your AI agent to run on a remote server but still be able to control your local machine? It’s a common hurdle when you want the 24/7 uptime of a VPS but need the agent to see your local screen or use your local camera.

How do commands propagate between Telegram, the gateway, and nodes?

The gateway is the brain of the operation. It handles your Telegram messages, runs the agent, and only reaches out to nodes via the Gateway WebSocket when a specific node tool is needed:

Telegram → Gateway → Agent → `node.*` → Node → Gateway → Telegram

Nodes are kept in the loop only for specific tasks; they don’t see the incoming provider traffic directly.

How can my agent access my computer if the Gateway is hosted remotely?

The trick is to pair your computer as a node. Your Gateway can live anywhere, but it can still call node.* tools like your screen, camera, or system commands on your local machine over the Gateway WebSocket.

Here is how you usually set it up:

1. Run the Gateway on an always-on host like a VPS or home server.
2. Put both the Gateway host and your computer on the same tailnet.
3. Ensure the Gateway WS is reachable via tailnet bind or an SSH tunnel.
4. Open the macOS app locally and connect in Remote over SSH mode.
5. Approve the node on the Gateway:

openclaw nodes pending
openclaw nodes approve <requestId>

Tailscale is connected but I get no replies. What now?

First, check the status of your components to see where the connection is dropping:

openclaw gateway status
openclaw status
openclaw channels status

If those look right, verify your auth and routing:

• If you use Tailscale Serve, check that gateway.auth.allowTailscale is set correctly.
• If you use an SSH tunnel, confirm the local tunnel is active and pointing at the right port.
• Confirm your allowlists (DM or group) include your account.
• Verify your network connection is stable.

Can two OpenClaw instances talk to each other (local + VPS)?

Yes. While there isn’t a built-in “bot-to-bot” bridge, you can connect them yourself.

The simplest way is to use a normal chat channel that both bots can access, like Telegram, Slack, or WhatsApp. You can also use the CLI:

openclaw agent --message "Hello from local bot" --deliver --channel telegram --reply-to <chat-id>

Just be sure to add a guardrail so the two bots do not end up in an infinite loop.

Do I need separate VPSes for multiple agents?

No. One Gateway can host multiple agents, each with its own workspace, model defaults, and routing. This is the standard setup because it is cheaper and easier to manage. You only need separate VPSes if you require strict isolation between agents.

curl -fsSL https://tailscale.com/install.sh | sh
sudo tailscale up

openclaw gateway --tailscale serve

openclaw nodes pending
openclaw nodes approve <requestId>

{
  agents: { defaults: { workspace: "~/.openclaw/workspace" } },
  channels: { whatsapp: { allowFrom: ["+15555550123"] } }
}