Summary

  • Several apps allow users to run local large language models (LLMs) on their devices, giving them offline access and stronger data privacy
  • Among the best are Ollama, Msty, AnythingLLM, Jan, and LM Studio
  • Ollama stands out for its simplicity and accessibility, supporting a variety of models and available on macOS, Windows, and Linux
  • To launch a model, users enter the command “ollama run [model identifier]” followed by the name of the supported LLM
  • Msty is easy to use and lets users avoid the command line, offering a library of prompts and workspaces to keep chats and tasks organised
  • AnythingLLM allows users to download models from third-party sources, including Ollama, LM Studio, and Local AI, enabling them to run thousands of LLM models available online
  • Jan and LM Studio provide easy, accessible ways to load popular LLM models like Llama, Mistral, Gemma, DeepSeek, Phi, and Qwen from Hugging Face

By Alvin Wanjala

Original Article