Local model training, inference, media generation, and agent workflows.
Local model inference
No code model training
Agent connections and API
Explore with AI
About
Unsloth Desktop is a free open source local AI workspace for running and training text, image, video, audio, embedding, and other models on personal hardware. It provides a desktop environment for downloading models, chatting locally, training supported models, generating media, and connecting developer agents to local inference. The product runs on Windows, macOS, and Linux and is designed for people who want a practical local model workflow without assembling every component separately.
Users can download and manage models, train supported models without writing code, generate images and video, run code in a sandbox, perform private web research, and connect Claude Code, Codex, and other agents to local models. An OpenAI compatible API lets applications and SDKs communicate with local inference, while optional remote access can expose models through a secure tunnel.
The main difference is the combination of local inference, training, media generation, agent connectivity, web research, and tool execution in one desktop application. The software is free and open source, but suitable hardware is still required and some models need substantial memory or compute. Training and inference capabilities vary by model and hardware, so users should check compatibility before committing to a local setup.
For FutureStack, Unsloth Desktop belongs in Developer Tools because its primary buyer intent is local model development and experimentation. It is best for running models locally, training models, connecting coding agents to local inference, building private AI workflows, and experimenting without sending every workload to a hosted provider.
We may earn commissions from links to support our work. Learn more.
Explore with AI
Pricing summary
Unsloth Desktop is free and open source. Users provide compatible hardware, while optional remote access or external infrastructure can introduce separate costs.
Pricing & Plans
Popular
Open Source
Free local software for running and experimenting with models.
Free
Local inference
Desktop workspace
Core model tools
Open source software
Self Hosted
Run models on your own compatible hardware.
Free
Own hardware
Local data control
Training workflows
Agent connections
External Infrastructure
Optional remote access and infrastructure.
Custom
Remote access
Cloud models
External infrastructure
Usage dependent costs
Other pricing notes
Pricing checked on 2026 08 24 from the official Unsloth website.
The desktop software is free and open source.
Hardware and optional infrastructure costs are separate.