🦞 OpenClaw Deploy

Deploy OpenClaw AI agents to Nebius Cloud.

→ Open the Deploy UI at claw.moi

Choose Your Path

Path	Method	Inference	Best For
1. Local Install	npm install -g openclaw	Token Factory	Try it now, zero overhead
2. Docker	docker run pre-built image	Token Factory	Portable, reproducible
3. GPU Serverless	NemoClaw on Nebius GPU	Local model	Custom models, data privacy
4. CPU Serverless	OpenClaw on Nebius CPU	Token Factory	Production, always-on

Quick Start

Path 1 — Install locally (30 seconds):

npm install -g openclaw
export TOKEN_FACTORY_API_KEY={your-key}
openclaw init && openclaw gateway --bind loopback --auth token

Path 2 — Docker (2 minutes):

docker run -e TOKEN_FACTORY_API_KEY={your-key} \
  -e TOKEN_FACTORY_URL=https://api.tokenfactory.nebius.com/v1 \
  -e INFERENCE_MODEL=zai-org/GLM-5 \
  -e OPENCLAW_WEB_PASSWORD={your-password} \
  -p 8080:8080 -p 18789:18789 \
  ghcr.io/colygon/openclaw-serverless:latest

Path 4 — Nebius CPU Serverless (3 minutes):

export TOKEN_FACTORY_API_KEY={your-key}
./install-openclaw-serverless.sh

Screenshots

Create Agent	Endpoints

What's Included

Deploy UI	Browser-based deployment wizard with endpoint management
Install Scripts	One-command deploy to Nebius serverless
Docker Images	Pre-built public images on GHCR
Setup Guide	Comprehensive Nebius configuration guide

Public Docker Images

docker pull ghcr.io/colygon/openclaw-serverless:latest   # ~400 MB
docker pull ghcr.io/colygon/nemoclaw-serverless:latest   # ~1.1 GB

Related

• nebius-skill — Claude Code skill for managing Nebius infrastructure
• OpenClaw — The open-source AI agent platform
• Token Factory — Nebius managed GPU inference API

License

MIT