Unsloth launches local studio for open model training

Unsloth has launched Unsloth Studio in beta, an open-source, no-code web UI for training, running and exporting open models locally. The tool supports GGUF and safetensor models on Mac, Windows and Linux, and can train 500+ models 2x faster with 70% less VRAM, according to Unsloth.

The interface supports text, vision, TTS audio and embedding models, with dataset creation from PDF, CSV, JSON, DOCX and TXT files. Users can run local models with web search, Bash and Python code execution, self-healing tool calling, automatic inference settings and API access for tools including Claude Code, Codex, Hermes, OpenClaw and OpenCode.

Unsloth Studio includes Data Recipes for turning documents into usable or synthetic datasets, training observability for loss, gradient norms and GPU utilization, side-by-side model comparison and export to safetensors or GGUF for use with llama.cpp, vLLM, Ollama and LM Studio. The beta works on Windows, Linux, WSL and MacOSx, with CPU support limited to chat inference and Data Recipes, while training works on NVIDIA, Intel, Apple and AMD GPUs.