AI on Ubuntu
Ubuntu supports a variety of AI[1] tools and workflows.
AI on Ubuntu Desktop
Is AI bundled into the Desktop?
As of Ubuntu 26.04 LTS, a fresh installation of Ubuntu Desktop contains no built-in AI tooling. So no AI features need to be configured or disabled.
Are there plans for AI features?
AI features are planned for future Desktop releases. These are focused on improving accessibility, and include voice-to-speech dictation[2], using local LLM models and a privacy-first design.
Developer toolchains
Develop ML models and AI applications on Desktop.
Python
Python is installed by default in recent versions of Ubuntu:
python3 --version
For more information, read about Python support in the official Ubuntu for developers documentation.
NVIDIA CUDA
From Ubuntu 26.04, installing the CUDA toolkit is straightforward:
sudo apt install cuda-toolkit
Local LLMs
Run LLMs on your Ubuntu machine without requiring a cloud service or subscription.
Inference snaps
These are models that you can install locally as a snap that are optimized for your machine.
Example:
sudo snap install gemma4
gemma4 chat
Read the official inference snap documentation.
Ollama
Run open models, locally on your machine or using Ollama's cloud models:
Example:
sudo snap install ollama
ollama run qwen3:0.6b
Read the Ollama docs.
Development environments
Workshop
Workshop is a container technology that can be used to make composable environments for safe agentic development.
Read the getting started tutorial in the official Workshop documentation.
Notes
- ↑ Here, the term “AI” is used broadly to refer to technologies that leverage generative AI, large language models, and AI agents. The term itself predates these developments by decades. Its usage here reflects its popular contemporary meaning, which is likely to be familiar to most readers.
- ↑ Introducing Myna: Speech to text for Ubuntu Desktop