Void with Ollama: local coding setup on Windows

This tutorial, uploaded on May 15, 2025, demonstrates a Windows setup for Void with Ollama. It does not specify software version numbers. The speaker presents Void as an open source Cursor alternative and a fork of VS Code. The download options described at the time cover Windows and Mac on Intel and ARM.

The walkthrough installs Void, selects its local provider option, then installs Ollama and downloads CodeGemma through PowerShell. Void connects to Ollama at localhost:11434 and detects the model. This gives viewers a concrete example of connecting a local LLM to a coding assistant, though the installation screens and provider options reflect the dated demonstration.

The demonstration machine has a GPU with 24 GB of VRAM. The speaker recommends that capacity for practical coding work and suggests a small model for CPU use. These are his recommendations, rather than established minimum requirements for Void.

Inside a local project, the speaker requests a codebase summary, adds selected code to chat with Ctrl+L, and applies suggested edits. He also asks for a Python program that reads URLs from a text file. A later debugging attempt fails, and he says the tool appears to have crashed. Starting a new chat lets him retry and accept a fix. He describes the Ollama workflow as private and local, but the tutorial does not independently verify data handling.