OpenClaw - RepoCloud Documentation

OpenClaw on RepoCloud

OpenClaw is an open-source AI coding platform that uses Anthropic’s Claude models to generate, test, and iterate on software in persistent workspaces. On RepoCloud, OpenClaw runs on a dedicated VPS with full root access, SSH, a browser console, and all the management tools in the RepoCloud dashboard.

Claude-Powered

Uses Anthropic’s Claude models for code generation, testing, debugging, and natural-language-to-code translation

Multi-Provider

Supports Anthropic, OpenAI, OpenRouter, and custom endpoints — switch models without code changes

Gateway Architecture

A secure gateway manages API credentials, token routing, and provider failover independently from your workspaces

Persistent Workspaces

Project files survive server reboots and sessions — pick up where you left off at any time

Full Root Access

SSH into your server, install custom toolchains, and manage the entire environment at the OS level

Your Infrastructure

All code, prompts, and API keys stay on your own server — nothing leaves unless you push it

Prerequisites

Before deploying OpenClaw on RepoCloud, you need an Anthropic API key:

Create an Anthropic Account

Go to console.anthropic.com and sign up or log in.

Add a Payment Method

Navigate to Billing and add a credit card. New accounts need a payment method before API keys will work.

Generate an API Key

Go to Settings > API Keys, click Create Key, and give it a name (e.g., “openclaw”). Copy the key immediately — it starts with sk-ant- and is shown only once.

OpenClaw uses pay-per-use Anthropic billing. The API key you provide powers all Claude calls; token costs are billed to your Anthropic account, separate from your RepoCloud subscription.

Deploying OpenClaw on RepoCloud

Find OpenClaw in the Marketplace

Go to the RepoCloud marketplace and search for “OpenClaw”. Click on the OpenClaw VPS listing.

Enter a Project Name

Choose a friendly name for your project (e.g., “my-openclaw”).

Enter Your Admin Email

Provide the email address for the OpenClaw admin account.

Enter Your Anthropic API Key

Paste your sk-ant-... API key. This is stored securely on the VPS and used by the OpenClaw gateway to authenticate with Anthropic’s API.

Choose Resource Settings

Select a tier for your server. Tiers below 2 GB RAM are greyed out because OpenClaw’s installation requires at least 2 GB. Servers can be upgraded to a higher tier but not downsized.

Deploy

A dedicated IPv4 address is included with every VPS at no extra charge. Click Deploy OpenClaw and wait 5–10 minutes.

Check Your Email

Once deployment finishes, you receive an email with:

Dashboard URL — the HTTPS link to your OpenClaw interface
App Credentials — gateway token for API access
SSH command — ssh root@<your-ip>
SSH Private Key — save it to a file, chmod 600, and use it to connect

First-Time Setup

Open the Dashboard

Open the dashboard URL from your deployment email. OpenClaw serves a web interface where you can interact with the AI assistant and manage workspaces.

Authenticate

Use the gateway token from your deployment email (also visible in the RepoCloud manage page under Access > App Credentials) to authenticate.

Verify Your API Connection

OpenClaw is pre-configured with the Anthropic API key you provided during deployment. You can verify the connection by starting a conversation — if the model responds, the key is working.

Configuring AI Models

Default Configuration

OpenClaw is deployed with Anthropic as the default provider. Your API key is stored in ~/.openclaw/.env on the server:

ANTHROPIC_API_KEY=sk-ant-your-key-here

The default model configuration lives in ~/.openclaw/openclaw.json:

{
  "env": { "ANTHROPIC_API_KEY": "sk-ant-..." },
  "agents": {
    "defaults": {
      "model": { "primary": "anthropic/claude-sonnet-4-6" }
    }
  }
}

Adding Additional Providers

You can add more AI providers by SSH-ing into the server and editing the configuration: OpenRouter (access to 400+ models with one key):

ssh root@<your-ip>
openclaw config set OPENROUTER_API_KEY sk-or-your-key-here

OpenAI:

openclaw config set OPENAI_API_KEY sk-your-key-here

Switching Models

Change the active model at any time:

openclaw config set model anthropic/claude-opus-4-8

Or use the interactive model selector:

openclaw model

This presents a menu of available providers and their supported models. The change takes effect immediately for new conversations.

Using OpenClaw

Code Generation

Start a conversation with a natural-language description of what you want to build. OpenClaw will:

Scaffold the project — create directory structure, config files, and boilerplate
Write implementation code — generate functions, classes, and modules based on your description
Create tests — write unit tests and integration tests alongside the implementation
Iterate on feedback — refine the code based on your review, error messages, or new requirements

The web interface shows diffs, terminal output, and the file tree in real time as the AI works.

Working with Existing Code

You can also point OpenClaw at existing projects:

Clone your repository onto the VPS via SSH
Open the project in OpenClaw
Ask the AI to refactor, add features, fix bugs, or write tests for existing code

Terminal and File Access

OpenClaw has direct access to the server’s file system and terminal. It can:

Run build commands (npm install, pip install, cargo build)
Execute tests and report results
Read and modify files across the entire project
Install system packages via apt

Security Considerations

Your Anthropic API key is stored on the VPS in ~/.openclaw/.env. Anyone with SSH access to the server can read it. Manage SSH keys carefully and remove keys for users who should no longer have access.

OpenClaw’s security model:

API keys stay on your server — they are never sent to RepoCloud or any third party
Code stays on your server — generated files live in the VPS file system; nothing is uploaded unless you push to a remote Git repository
Network isolation — the gateway process communicates only with the configured AI provider APIs (e.g., api.anthropic.com)
File permissions — ~/.openclaw/.env should be chmod 600 (this is the default)

To rotate your API key:

ssh root@<your-ip>
openclaw config set ANTHROPIC_API_KEY sk-ant-new-key-here
openclaw gateway restart

RepoCloud Manage Page

Your OpenClaw VPS has a dedicated manage page in the RepoCloud dashboard with these sections:

Access

View your server’s default domain (HTTPS link to OpenClaw), dedicated IPv4 address, SSH connection command, and app credentials (gateway token).

Power

Power Off, Power On, and Reboot controls.

Console

Browser-based serial console for emergency access.

SSH Keys

Manage SSH public keys for root access.

Resize

Switch between tiers. Disk cannot be shrunk after an upgrade.

Snapshots

Create up to 5 point-in-time snapshots before making changes. To restore, click Restore next to the target snapshot. If newer snapshots exist, a warning lists which ones will be permanently destroyed. The server restarts during restore (~10–25 seconds) and the page refreshes automatically when complete.

Custom Domain

Point your own domain to the VPS by creating an A record to your IPv4.

Danger Zone

Permanently delete the server and all data.

Troubleshooting

'No credentials found' error

The Anthropic API key is not configured or the gateway cannot read it. SSH into the server and check: cat ~/.openclaw/.env | grep ANTHROPIC. If the key is missing, re-add it with openclaw config set ANTHROPIC_API_KEY sk-ant-... and restart the gateway.

Model responses are slow or timing out

Large prompts or complex code generation tasks can take 30–60 seconds. If responses consistently time out, check your Anthropic account for rate limits or billing issues. You can also try a faster model (e.g., anthropic/claude-sonnet-4-6 instead of anthropic/claude-opus-4-8).

Installation fails with out-of-memory

OpenClaw requires at least 2 GB RAM. The installation process pulls Docker images and compiles dependencies; on 1 GB servers, this causes OOM kills. Resize to 2 GB or larger from the RepoCloud manage page.

Cannot access the dashboard

Wait 10–15 minutes after deployment for the installation to complete. If it remains unreachable, use the browser console from the RepoCloud manage page to check service status. Verify that the gateway process is running.

API key charges seem high

OpenClaw sends your code context to Claude on each interaction. Large codebases increase token usage. Use the Anthropic console to monitor usage and set spend limits. Consider using a smaller model for routine tasks and reserving larger models for complex problems.

Use Cases

OpenClaw is ideal for:

Solo developers who want an AI pair programmer on their own infrastructure with no data leaving their server
Teams that need a shared AI coding environment with persistent workspaces and centralized API key management
Security-conscious organizations that require all code and prompts to stay on controlled infrastructure
Developers exploring AI coding who want a self-hosted alternative to cloud-based AI editors

Next Steps

Deploy OpenClaw

Launch your OpenClaw VPS from the RepoCloud marketplace

OpenClaw Documentation

Read the official OpenClaw documentation

Hermes Agent

Explore Hermes Agent for autonomous multi-model AI workflows

VPS Apps Overview

Learn about all VPS Apps and the RepoCloud manage page

​OpenClaw on RepoCloud

Claude-Powered

Multi-Provider

Gateway Architecture

Persistent Workspaces

Full Root Access

Your Infrastructure

​Prerequisites

​Deploying OpenClaw on RepoCloud

​First-Time Setup

​Configuring AI Models

​Default Configuration

​Adding Additional Providers

​Switching Models

​Using OpenClaw

​Code Generation

​Working with Existing Code

​Terminal and File Access

​Security Considerations

​RepoCloud Manage Page

​Troubleshooting

​Use Cases

​Next Steps

Deploy OpenClaw

OpenClaw Documentation

Hermes Agent

VPS Apps Overview

OpenClaw on RepoCloud

Prerequisites

Deploying OpenClaw on RepoCloud

First-Time Setup

Configuring AI Models

Default Configuration

Adding Additional Providers

Switching Models

Using OpenClaw

Code Generation

Working with Existing Code

Terminal and File Access

Security Considerations

RepoCloud Manage Page

Troubleshooting

Use Cases

Next Steps