Unsloth Desktop Launches: Train & Run 500+ Local Models, 70% Less VRAM
danielhanchen · x · 2026-09-17
Unsloth released Unsloth Desktop, the first open-source desktop app to run and train models locally (Mac/Windows/Linux), plus an official Docker image supporting 500+ models with no setup on NVIDIA and AMD GPUs.
- Trains models 2× faster with 70% less VRAM; supports MLX, diffusion image/video, audio, GGUF
- Connect Claude Code and Codex to local LLMs; sandboxed code exec, private web search, RAG, MCP
- OpenAI-compatible API and secure remote deployment
- Docker exposes ports 8000 (Studio) and 8888 (JupyterLab); needs NVIDIA driver 570.26+
Repo, blog and install guide are public.
Related event: Unsloth Desktop Brings Local Training of 500+ Models(2 posts)→
More from coding & agent
- Video editor claims Claude cut her entire video via text prompts, no editing software — CodeByPoonam · 2026-11-02
- Cohere Labs agent team-construction method cuts multi-agent costs by 70% — Cohere_Labs · 2026-11-02
- Long-read: AI's real value comes after the efficiency era—infra glut, agent tooling boom — yangyi · 2026-11-02
- offprompt: MIT-licensed MCP server hands agents API keys without ever exposing them — Vast_Ad_7338 · 2026-11-02
- Reddit idea: expose every OOP public method as AI agent tools to agentify legacy apps — out_of_nowhere__ · 2026-11-02
- DeepSeek design doc: a person can only follow 3-5 agent sessions at once, attention is the real bottleneck — heyshrutimishra · 2026-11-02