Player not loading? Watch on YouTube
This tutorial walks through Pinokio AI setup and uses an Ollama chatbot to demonstrate how to run models locally. The presenter selects the Apple Silicon download for macOS; the download page also lists Windows and Linux. Initial settings include an installation location, theme, and desktop or background mode. The speaker advises using the internal drive rather than an external drive.
The Explore tab provides filters for platform, GPU vendor, app type, and tags. After selecting the Ollama chatbot and a branch, the presenter starts an installation that lists 13 requirements. Pinokio installs these automatically in the demonstration, with Conda shown in progress. An optional protection setting comes with a warning that some installations may fail; the presenter leaves it off.
The next step requires Ollama to be installed and open. The presenter downloads a model identified in the transcript as "3.1 latest," then launches the chatbot through a localhost address. A greeting and a request about elephants demonstrate responses from the local LLM. The interface also has a control for response randomness and options to create chats, import or export data, and clear conversations.
The tutorial ends by stopping the chatbot in Pinokio and explaining how to start it again. Downloads require an internet connection in this walkthrough.
