Player not loading? Watch on YouTube
This overview introduces five workflows for running AI tools on your own computer, with brief setup instructions and examples. The examples cover local chat, web research, a code fix, dictation and image generation.
Jan is the local LLM chat app in the walkthrough. The speaker downloads it for Windows or Mac, selects the recommended Jan V3.5 model and tests follow-up prompts. He says it can work offline after setup and judges its responses similar to paid chatbots.
Vane handles internet research through a Docker setup. Its demonstration covers speed and quality modes, answers with linked sources, searches restricted to chosen websites, uploaded files and locally saved history.
OpenCode is a coding assistant that reads project files and applies edits. In the example, it corrects a discount calculation that returns 199.9 instead of 180, asking permission before changing and testing the code. WhisperX follows as a speech-to-text option, with Python installation required before installing the tool.
For images, the speaker installs InvokeAI and selects FLUX.2 Klein 4B through its model manager. He recommends the smaller Klein model when hardware struggles with the heavier model. The prompt-writing tool called Vision Prompt Studio in the description supplies a structured prompt; the speaker says access requires Gmail. Image generation takes a couple of minutes in the example, and performance depends on the computer.
