Running OpenClaw on Nebius

About this workshop

Colin Lowenberg and Michal from Nebius hosted a webinar demonstrating how to run OpenClaw on Nebius' cloud infrastructure, specifically using Token Factory and Serverless AI services. The presentation covered installing OpenClaw locally and connecting it to Token Factory's API for accessing various open-source language models at lower costs compared to proprietary services like ChatGPT. Michal explained how to deploy OpenClaw to Nebius Serverless using Docker containers, showing the process through a CLI skill that automates the deployment. The session included live demonstrations of configuring different models, setting up sub-agents, and managing security considerations. Participants asked questions about phone number integration, model switching, cost estimation, and deployment from GitHub, with Michal confirming that GitHub deployment is on their roadmap. The webinar concluded with discussions about AI's impact on jobs and recommendations for using Tavily for web search capabilities within OpenClaw.

Recommended reading

• opencolin/openclaw-deploy — local install, Docker, Nebius CPU & GPU paths
• docs.contree.dev — sandboxed workspaces with Git-like branching
• Token Factory docs — model catalog & cost patterns