01

Image and Video Generation

Create images with fine-tuned LoRA adapters and generate video with Wan2.2-TI2V-5B and LTX-2.3 GGUFs.

Transform, inpaint, extend, upscale, reference and edit existing images - all locally.

Local image generation in Unsloth Desktop
02

Use the latest models

Run and train nearly every model locally, with Day Zero support for releases like Qwen3.8, NVIDIA, GLM, Gemma and more.

Discover, manage and download the right quantization for your device from the built-in model hub.

Unsloth Desktop model hub
03

Models that actually run code

Most apps stop at JavaScript. Unsloth Desktop runs Bash and Python in a sandbox, so models can test their work instead of guessing.

Broken tool calls are repaired rather than dropped.

Sandboxed code execution in Unsloth Desktop
04

Advanced Web search

Normal search uses the web while your model is still thinking.

Deep Research plans first, finds the best sources, then creates a detailed report with citations.

Advanced web search in Unsloth Desktop
05

Access models anywhere

Serve local or Colab models over HTTPS through a free Cloudflare tunnel.

Check a run from your phone, laptop, or anywhere else you happen to be.

Secure remote access to Unsloth Desktop
06

Connect your Agent

Connect Claude Code, Codex and other agents to local models with the unsloth start command.

Start Unsloth, load a model, open your project folder, then run unsloth start claude.

Connect Claude Code to Unsloth Desktop
Unsloth Desktop

Your complete local AI workspace

Run, train and connect the latest text, image, video and audio models on Mac, Windows and Linux. Free, open-source and built to keep your work local.

01

Download a model and start chatting in minutes.

02

Train and generate media with no-code workflows.

03

Connect Claude Code, Codex and more to your own GPU.

Open source and free

Ready to use Unsloth?

New
✨Introducing Unsloth Studio

Meet Unsloth Studio

Meet Unsloth Studio

Meet Unsloth Desktop

The first desktop app to run and train AI models.
Open-source. Free. 100% Local.

View more news

Run models locally

Unsloth Studio runs 100% offline on your Mac and Windows device. Run GGUF and Safetensors models with tool-calling, web search, and OpenAI compatible API.

Compare models side by side and upload images, docs, audio, code files and more.

Run models locally on Mac, Windows, Linux
Train models with no-code and full observability.

No-code training

Auto-create datasets from PDF, CSV, JSON docs and start training with real-time observability.

Unsloth's custom kernels supports optimized training for LoRA, FP8, FFT, PT and 500+ models including text, vision, audio and embeddings.

Unlimited Tool calling + Web search

Unsloth Studio lets LLMs run unlimited web search and execute Bash and Python, not just JavaScript. It sandboxes programs like Claude Artifacts so models can test code, generate files + verify answers with real computation.

E.g. Qwen3.5-4B searched 20+ websites and cited sources, with web search happening inside its thinking trace.

Use Unsloth Data Recipes to create and customize datasets from files.

Data Recipes

Data Recipes transforms your docs into useable datasets via graph-node workflow. Upload unstructured or structured files like PDFs, CSV and JSON. Unsloth Data Recipes auto turns documents into your desired formats.

Export models

Export any model, including your fine-tuned models, to safetensors, or GGUF for use with llama.cpp, vLLM, Ollama, and more.

Export models with Unsloth Studio

Don’t believe us?

Why not try our fully free open source version? Finetune 2X faster on a single NVIDIA GPU for free on Google Colab or Kaggle Notebooks.

Sign up to our newsletter
We'll share monthly updates!
Subscribe now

Train your own custom model in 24 hrs, not 30 days.

30x faster than FA2 + 30% accuracy
90% less memory usage than FA2
audio, embedding, vision support

The details

We're making AI more accessible to everyone

Find out more

Unsloth makes everything greener

As hardware costs rise and performance gains plateau, we use our math and coding skills to make models train and run smarter + faster.

Want lightning fast inference? We’re working on it!

Contact us

Don't forget to join our newsletter!

By registering you agree to unsloth's Terms of Service and Privacy Policy,
Subscribe now
MultiGPU Docs

Even better multiGPU in the works!

Don't forget to join our newsletter!

Subscribe

Ready to use unsloth?

Join Our Discord