Build an AI Assistant Without OpenAI — Fully Private and Free to Run

This workshop uses Docker Model Runner to run open weight models including Llama, Mistral, Phi, and Gemma locally on your machine. The instructor covers how to select and configure the best model for your hardware and use case — with no OpenAI account or API key required at any point.

Yes. Modern open weight models available through Docker Model Runner are excellent for real personal assistant tasks including question answering, writing assistance, summarisation, coding help, and conversational interaction. For the majority of everyday use cases they are very capable alternatives to OpenAI models.

It used to be. Docker Model Runner and OpenClaw in 2026 have made it straightforward. This live workshop guides you through the complete setup in 4 hours — far faster than piecing it together from documentation on your own.

Not necessarily. The quality gap between frontier models and the best open weight models has narrowed significantly. For personal assistant tasks the difference is minimal. This workshop helps you choose the right model and configure OpenClaw to get the best possible responses from your local setup.

Yes, periodically. The advantage of building your own AI assistant without OpenAI is that you control when and how to update. You can pull a newer model through Docker Model Runner whenever you want — with no forced upgrades, no pricing changes, and no terms of service updates.

Yes. Your private AI assistant built in this workshop runs entirely on your own machine with no data leaving your hardware. This makes it suitable for sensitive business tasks where sending data to OpenAI's servers would be a concern. The instructor covers security configuration during the live session.