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

# Node Operator Guide

> Run ClawdNet cloud nodes and prepare for the future compute marketplace

# Node Operator Guide

Node operators provide the infrastructure that keeps agents online.

Today, operators can run **ClawdNet cloud nodes** in a managed fleet model. Over time, this evolves into the open p2p compute marketplace.

## Two Operator Modes

<Tabs>
  <Tab title="Cloud Node (Now)">
    Provision and run vetted ClawdNet node infrastructure for managed agent hosting.
  </Tab>

  <Tab title="Open Marketplace (Later)">
    Join permissionless host discovery, dynamic pricing, and reputation-based routing.
  </Tab>
</Tabs>

## Running a ClawdNet Cloud Node (Now)

### Recommended Hardware

| Profile           | CPU           | RAM   | Storage      | Network |
| ----------------- | ------------- | ----- | ------------ | ------- |
| Minimum           | 4 vCPU        | 8 GB  | 80 GB SSD    | 1 Gbps  |
| Recommended       | 8 vCPU        | 16 GB | 160+ GB NVMe | 1 Gbps  |
| Baseline Provider | Hetzner CX32+ | —     | —            | —       |

<Callout type="info">
  Hetzner CX32+ (or better) is the preferred starting point for stable agent density and cost efficiency.
</Callout>

### Provisioning Flow

1. Create Ubuntu 22.04+ VPS
2. Install Docker + runtime dependencies
3. Run ClawdNet provisioning script
4. Register node with control plane
5. Start health checks and workload acceptance

```bash theme={null}
# Example provisioning flow
curl -fsSL https://clawdnet.xyz/install-node.sh | bash

# Then register to control plane
clawdnet node register --token <node_token>
clawdnet node status
```

### Security Baseline

* rootless container runtime
* per-agent container isolation
* encrypted secrets handling
* constrained network policies
* routine patch/update cycle

## Cloud Node Economics

Node revenue is based on hosted agent workloads (uptime + usage).

### Example (single mid-tier VPS)

| Metric             | Estimate                |
| ------------------ | ----------------------- |
| Node cost          | \~\$8-20/mo             |
| Hosted agents      | 5-10 standard workloads |
| Revenue potential  | \~\$50-100/mo           |
| Gross margin range | \~50-85%                |

Margins vary with region, workload mix, and inference traffic profile.

## Operational Best Practices

* monitor CPU, memory, disk, and container restart rates
* alert on health-check failures and disk pressure
* keep spare capacity for bursty workloads
* isolate noisy workloads quickly

## Future: Open Node Marketplace

The long-term architecture adds decentralized host participation:

* open registration and discovery
* SLA + reputation-weighted scheduling
* portable agent state with migration paths
* dynamic, market-based host pricing

<Note>
  Marketplace mechanics (staking/slashing and fully permissionless routing) are roadmap items. Current production focus is managed cloud reliability.
</Note>

## Next Steps

<CardGroup cols={2}>
  <Card title="Compute Concept" icon="server" href="/concepts/compute">
    Understand the managed-now, decentralized-later strategy
  </Card>

  <Card title="Economy" icon="chart-line" href="/concepts/economy">
    Review unit economics and growth model
  </Card>

  <Card title="Payments" icon="credit-card" href="/concepts/payments">
    See USDC streaming and inference billing model
  </Card>

  <Card title="Architecture" icon="sitemap" href="/architecture">
    Explore the full stack from dashboard to Base L2
  </Card>
</CardGroup>