Running an OpenClaw backend on your VPS
read · 7 min
OpenClaw is open-source software — the platform is free, you only pay for the model you point it at. Hosting it yourself means your assistant runs 24/7 and you control every layer, from model to messaging channel.
Steps
Install Docker and Compose
OpenClaw ships as containers. Install Docker Engine and the Compose plugin.
$ curl -fsSL https://get.docker.com | sh
$ sudo docker compose version Clone the repository
Pull the source onto your VPS.
$ git clone https://github.com/openclaw/openclaw.git
$ cd openclaw Configure your environment
Copy the example env file and fill in your model provider key. Pick any backend — Claude Opus, GPT, Gemini, DeepSeek or a local Ollama endpoint.
$ cp .env.example .env
$ nano .env # set MODEL_PROVIDER and API key Start the stack
Bring the containers up in the background. They restart automatically after a reboot.
$ sudo docker compose up -d
$ sudo docker compose logs -f Connect a messaging channel
Add a channel token (Telegram, WhatsApp, Discord) in your .env and restart. Your assistant is now reachable from your phone.
$ sudo docker compose restart Frequently asked
Is OpenClaw free? +
The software is open-source and free. Your only cost is the model — a flat VPS fee plus per-token API usage, or zero extra if you run a local model.
Which models can it use? +
Any major provider: Claude, GPT, Gemini, DeepSeek, or a local model served through Ollama. You set the provider in .env.
Can it handle several channels at once? +
Yes — Telegram, WhatsApp, Discord and others can run side by side from the same backend.
Related guides
Ready to self-host your assistant?
Get a VPS with the resources and uptime an always-on assistant needs.
See VPS plans →