Advanced AI Build Night
Advanced AI Build Night · Local Models
Own Your AI
Run a capable open-weight model on your own machine with Ollama, point an OpenWork agent at your own files, and keep your data yours — no cloud required.
In this playbook
- The presentationRun the slides full-screen. Use ← / → or Space to advance.
- Cheat sheetThe whole topic on one printable page.
- Before you startAccounts and tools to have ready before the session.
- OllamaThe engine — installs in a click and serves models on localhost:11434.
- OpenWorkThe point-and-click cockpit — an open-source desktop agent that drives the model over your files.
- Ollama model libraryBrowse current open-weight models, sizes, and tags before you pull one.
- OpenWork → Ollama setupThe exact config to point OpenWork at your local Ollama endpoint.
- Ollama CloudRun big models like glm-5.2:cloud through the same CLI — datacenter GPUs, no local RAM.
- Claude Code on OllamaPoint Anthropic's Claude Code CLI at a local model — `ollama launch claude`.
- Transformers.jsRun models in the browser via WebGPU — the WebAI starting point.
- Unsloth — fine-tuningThe fastest, lowest-memory way to LoRA/QLoRA a small model on your own data.