What Ollama said on X

On March 26, 2026, Ollama said Visual Studio Code now integrates with Ollama through GitHub Copilot, so any local or cloud model available in Ollama can be selected directly inside VS Code. That matters because the integration is not framed as a side extension or a separate chat panel. It moves Ollama models into the same model picker many developers already use for Copilot-powered workflows.

What the docs add

Ollama’s VS Code integration page says this path requires Ollama v0.18.3 or newer, VS Code 1.113 or newer, and GitHub Copilot Chat extension 0.41.0 or newer. It also says users still need to sign in to use the model selector, but a paid GitHub Copilot subscription is not required; GitHub Copilot Free is enough to enable custom model selection.

Setup is deliberately lightweight. Ollama documents a one-command flow, ollama launch vscode, plus a manual path where users open Copilot Chat, bring up the Language Models window, add Ollama as a provider, and unhide the models they want. The docs also explicitly support cloud models, not just local weights, which means the same VS Code surface can bridge laptop-resident models and larger hosted ones.

Why it matters

The practical shift is that local and self-directed model choice is becoming a first-class workflow in a mainstream editor instead of an enthusiast setup. Developers can stay inside GitHub Copilot Chat while choosing Ollama-served models for privacy, cost control, or open-model experimentation, then fall back to hosted models when they need more scale. That lowers switching cost between closed and open ecosystems and gives local AI tooling a much more normal path into day-to-day software work.

Source: Ollama X post · Ollama VS Code docs

#vscode

Ollama models arrive in VS Code's GitHub Copilot Chat picker