Collection
A collection of neural networks and AI tools you can run locally on your computer or your own server.
A collection of neural networks and AI tools you can run locally on your computer or your own server.
18+ Telegram bot for animating photos into short videos
18+ Telegram bot for animating photos into short videos
18+ Telegram bot for AI photo editing and clothing-removal transformations
Yes, if the application and model are already installed on your device. You usually need internet access at the start to download the app, model weights, and updates. After that, local chat and generation can work offline.
For a first local setup, LM Studio, Jan, GPT4All, or Msty Studio are usually the easiest options. They offer desktop interfaces, model selection, and less manual configuration than server-based solutions.
For API use, people often choose Ollama or LocalAI. Ollama is easier to get started with and works well for developers, while LocalAI is useful when you need an OpenAI-compatible backend and more control over a self-hosted stack.
Not always. Smaller LLMs can run on a CPU, but performance will be slower. For larger models and image generation, a modern GPU or Apple Silicon with enough memory is preferable.
GPT4All and AnythingLLM are convenient for chatting with local files. LM Studio also supports working with documents offline. For a team-oriented self-hosted option, Open WebUI is worth considering.
Yes. DiffusionBee is a simpler option for macOS, while ComfyUI and other interfaces built around Stable Diffusion offer more advanced configuration. As with text models, you need to download the models and dependencies first.
Browse other collections — the right tool may already be there.
Browse collections
