GPT4All

GPT4All is one of the original open-source local LLM desktop apps, maintained by Nomic AI. It bundles a polished native UI for macOS, Windows, and Linux with a curated model catalog (Llama 3, Qwen, Mistral, Phi, DeepSeek, Nous Hermes, and many quantized community models) so that anyone can download a model, double-click it, and start chatting with no setup. Inference runs on CPU or GPU via llama.cpp under the hood, with sensible defaults that work even on modest laptops. Beyond chat, GPT4All ships LocalDocs — a built-in retrieval layer that lets you point the app at folders of PDFs, Markdown, code, or other text and ask questions against your own documents without any data leaving the device. There is also a Python and a Node.js SDK for embedding GPT4All into apps, and a HTTP server mode so the desktop app can act as a local OpenAI-compatible endpoint for other tools. The base product is free and open source under MIT; Nomic offers paid Atlas/Embedding cloud services for teams that need hosted embeddings, dataset visualization, and enterprise document workflows. GPT4All is widely used as an on-ramp for people new to local LLMs, by privacy-sensitive professionals (legal, medical, defense), and as a simple way to give friends and family a fully private AI chat experience.