AnythingLLM

About

AnythingLLM is a free and open-source, privacy-focused, multi-user ChatGPT application developed in the United States. It is designed to work with any Large Language Models (LLMs) and vector databases, offering unlimited documents, messages, and storage. Available as a desktop application for Mac and Windows, and also self-hostable via Docker, its core features include AI Writing, AI Chatbot, and Vector Database capabilities. The platform supports a wide range of LLMs such as localai, Ollama, Mistral-7b, Llama, and OpenAI ChatGPT, and integrates with vector databases like Milvus, Pinecone, and Chroma, utilizing Retrieval-Augmented Generation (RAG) and Langchain. Users can suggest and vote on new features. While generally well-received, with one user noting it's "Good as Openrouter desktop app," another user observed that an internet connection is required for document embedding when using its built-in embedding model. The project's source code is hosted on GitHub, and it maintains a presence on Twitter.