NovitaClaw: Run OpenClaw in the Cloud with One Command

Andrej Karpathy recently described a shift worth paying attention to: just as LLM agents emerged as a new layer on top of LLMs, Claws are the next layer on top of agents, taking orchestration, scheduling, context, tool calls, and persistence further than agents alone ever could.

That layer needs infrastructure that stays on. For developers running OpenClaw on Novita Sandbox, that meant working around session limits until now.

NovitaClaw closes that gap. It’s a new deployment tool on Novita Agent Sandbox that provisions a fully configured OpenClaw instance in a single command, with no runtime cap, no manual restarts, and no lifecycle management overhead.

The Problem the CLI Solves

Novita Agent Sandbox has supported OpenClaw deployment through the Sandbox Template. That path is suitable for exploration, but it carries a hard constraint: sandboxes spun up through the UI have a session lifecycle limit. After a few hours, the sandbox stops, and the agent goes offline. Developers who needed continuous availability had to work around this themselves — scheduling pauses, scripting restarts, or manually reviving instances.

NovitaClaw works differently. There is no session cap. The instance runs 24/7 for as long as you need it. The default spec — 2 vCPU, 4 GB RAM — is built for real workloads, not demos. And because the model layer is fully configurable, you choose which LLMs power your agent, whether hosted on Novita or sourced from a third-party provider.

Getting Started

Prerequisites

• Python installed
• A Novita API key — here’s how to get one:
  • Log into novita.ai — sign in with Google or GitHub (a new account is created automatically on first login), or sign up with your email address
  • Create an API key — go to the Key Management settings page to create or manage your API keys. Copy it somewhere handy — you’ll need it in the next step.

python3 --version

sudo apt update && sudo apt install python3 python3-pip -y

Install and Launch

Step 1: Install NovitaClaw

macOS / Linux:

sudo pip3 install novitaclaw

Windows PowerShell:

pip install novitaclaw

If you have a previous version installed, upgrade to the latest version:

Bash

pip3 install novitaclaw --upgrade

If the upgrade fails, try a force reinstall:

Bash

pip3 install novitaclaw --upgrade --force-reinstall

After installation, verify it by typing novitaclaw --help. If you see a list of instructions, you’ve succeeded!

Step 2: Set the environment variable in your terminal

macOS / Linux:

export NOVITA_API_KEY=sk_your_api_key

Windows PowerShell:

$env:NOVITA_API_KEY = "sk_your_api_key"

Step 3: Launch instance

novitaclaw launch

On success, the CLI returns three values you’ll use to access and manage your agent:

• Web UI URL
• Gateway WebSocket URL & Token
• Web Terminal URL (for terminal access to the sandbox)
• File Manager URL (for browsing and managing workspace files)
• Login credentials (for Web Terminal & File Manager)

Open the returned Web UI URL and go to the Chat tab to use your agent. Use the Web Terminal URL to open a terminal session inside the sandbox, and the File Manager URL to browse and manage files in the sandbox workspace.

Configuring Models

Your instance is pre-configured with a Novita-hosted model out of the box.

To change the models your agent uses, navigate to Settings → Config, click Raw to switch to Raw JSON5 view, then click the reveal button next to “2 secrets redacted” to show the full config. Update the following two sections:

Step 1: Register the model under your provider

Add a new object to the models array inside models.providers.novita:

"models": [
  {
    "id": "model-id",
    "name": "display name",
    "reasoning": true,
    "input": ["text"],
    "contextWindow": 200000,
    "maxTokens": 50000
  }
]

Step 2: Set it as primary or fallback

Update the model field under agents.defaults to reference your model using the provider/model-id format.

"agents": {
  "defaults": {
    "model": {
      "primary": "provider/model-id",
      "fallbacks": [
        "provider/model-id"
      ]
    }
  }
}

Click Update to save. Every LLM available on the Novita platform is supported. Third-party providers can also be configured — when you bring your own LLM, you pay only for sandbox runtime, not Novita model usage.

Your Claw is running. Now make it smarter.

BotLearn is the world’s first Bot University — a learning and distribution layer built natively on top of OpenClaw. Bots learn, humans earn: your agent gets structured Skills, benchmarking, and access to Botcord, the Agent-to-Agent marketplace where agents earn real money completing tasks.

To enroll your agent, just send it this message:

“This is the Bot University I want you to join. Read https://www.botlearn.ai/SKILL.md and follow the instructions to join BotLearn. If you encounter any security concerns during installation, please confirm each step with me before proceeding.”

Your agent handles the rest.

Connecting Channels

OpenClaw supports external messaging channels so your agent is reachable outside the Web UI. Channels require configuration — they are not active by default.

Telegram

Step 1: Create a Telegram Bot

1. Open Telegram and search for @BotFather
2. Send /newbot and follow the prompts to name your bot
3. Copy the bot token BotFather provides

Step 2: Send the bot token to Chat tab

In the Web UI, go to the Chat tab and send the bot token.

Step 3: Pair and use Telegram Bot

In Telegram, open the bot you created and send /start. Copy the code and paste it in the Chat tab.

Once connected, messages sent to your Telegram bot are routed directly to your OpenClaw agent. More channel integrations are supported — see the official OpenClaw channels guide for the full list and setup instructions.

Service Reliability

All core services in the sandbox are managed by systemd for production-grade reliability:

Crash auto-recovery: If the Gateway crashes repeatedly, the system automatically runs diagnostics, attempts repair, and restores the last known-good configuration from backup — no manual intervention required.

Config auto-backup: Every configuration write creates an automatic backup. If a bad config causes a crash, the recovery process restores from the most recent valid backup.

Command Reference

launch

Create a new sandbox instance.

Bash

novitaclaw launch [OPTIONS]

​list

List active sandbox instances.

Bash

novitaclaw list --api-key YOUR_API_KEY

​status

Check a sandbox instance status.

Bash

novitaclaw status <SANDBOX_ID> --api-key YOUR_API_KEY

​stop

Terminate a sandbox instance.

Bash

novitaclaw stop <SANDBOX_ID> --api-key YOUR_API_KEY

​tui

Connect to a sandbox via terminal UI.

Bash

novitaclaw tui <SANDBOX_ID> --token <GATEWAY_TOKEN>

​doctor

Run diagnostics and repair checks on a sandbox instance.

Bash

novitaclaw doctor <SANDBOX_ID> --api-key YOUR_API_KEY

​gateway update

Update OpenClaw to the latest version inside a sandbox.

Bash

novitaclaw gateway update <SANDBOX_ID> --api-key YOUR_API_KEY

​gateway restart

Restart the OpenClaw Gateway in a sandbox.

Bash

novitaclaw gateway restart <SANDBOX_ID> --api-key YOUR_API_KEY

​services setup

Install and start Web Terminal and File Manager on an existing sandbox. Use this to add services to older sandboxes that were launched before these services were included.

Bash

novitaclaw services setup <SANDBOX_ID> --api-key YOUR_API_KEY

​pair list

List pending pairing requests for a channel.

Bash

novitaclaw pair list <SANDBOX_ID> --channel <CHANNEL_NAME> --api-key YOUR_API_KEY

​pair approve

Approve a pairing request for a channel.

Bash

novitaclaw pair approve <SANDBOX_ID> --channel <CHANNEL_NAME> --code <PAIRING_CODE> --api-key YOUR_API_KEY

​update

Check for CLI updates and upgrade to the latest version.

Bash

novitaclaw update

​Programmatic Usage

Use JSON output for automation:

Bash

novitaclaw -j launch --api-key YOUR_API_KEY

Example response:

{
  "success": true,
  "data": {
    "sandbox_id": "xxx",
    "webui": "https://xxx.sandbox.novita.ai?token=xxx",
    "gateway_ws": "wss://xxx.sandbox.novita.ai",
    "gateway_token": "xxx",
    "terminal_url": "https://7681-xxx.sandbox.novita.ai",
    "filemanager_url": "https://7682-xxx.sandbox.novita.ai",
    "services_username": "admin",
    "services_password": "xxx"
  }
}

Environment Variables

Pricing

Troubleshooting

​Command not found

Try one of these:

Bash

# Use full path (Linux)
~/.local/bin/novitaclaw launch

# Typical macOS user path
~/Library/Python/3.x/bin/novitaclaw launch

​Connection timeout

Increase launch timeout:

Bash

novitaclaw launch --timeout 180 --api-key YOUR_API_KEY

Frequently Asked Questions

Build the Next Layer

The shift Karpathy describes — from agents to Claws — isn’t theoretical. It’s an architectural opportunity available right now, and the teams who build on persistent, always-on infrastructure will be the ones who realize it.

NovitaClaw gives you that foundation: a production-grade agent runtime, deployed in under a minute, with full control over the model layer and channel integrations. The sandbox handles uptime. The CLI handles provisioning. What you build on top is up to you.

Start building on Novita Agent Sandbox →

Novita AI is an AI and Agent Cloud for builders and startups — providing access to 200+ LLMs, image, video, and TTS models via one API, alongside GPU instances and Agent Sandboxes that launch in seconds.