Unsloth Launches Free Desktop Beta for Local AI Training

Open-source AI startup Unsloth is bringing model inference and fine-tuning to personal computers with Unsloth Desktop, targeting developers seeking greater control over data and computing resources. Running models locally can reduce reliance on cloud infrastructure and help keep sensitive information on-device, while simplifying experimentation with large language models across development workflows.

As of Aug. 13, 2026, Unsloth has released Unsloth Desktop Beta at no cost, with direct connections to AI coding tools including Claude Code and Codex. The software also supports web search, code execution in isolated environments and GPU computing across multiple platforms, combining local inference and training functions in a single desktop product aimed at lowering technical and financial barriers.