GPT4All

GPT4All is an open-source desktop ecosystem for developers, teams, and AI power-users who want to run large language models locally on Windows, macOS, and Linux.

Key capabilities include:

• On-device LLM inference on consumer hardware (CPU-friendly), designed to work without cloud services, API calls, or a GPU.
• Desktop app for chat and model management, with support for downloading and switching between thousands of open-source models (including families such as Llama 3, Mistral, Phi 3, and Gemma).
• LocalDocs for local document chat (RAG), which integrates local files into conversations using Nomic embedding models so data stays on the machine.
• Developer access via a Python client, leveraging llama.cpp backends for programmatic workflows and assistants.

Pricing: Free (open-source).

What makes it notable is its focus on privacy-first, offline-capable local AI, combined with an integrated model catalog, LocalDocs retrieval, and developer-friendly tooling that avoids dependence on hosted LLM APIs.