AnythingLLM

All-in-one private AI app for chatting with your documents

Alternative to: chatgpt, notebooklm

AnythingLLM is an all-in-one AI application that lets you chat with your documents, run AI agents, and connect to any local or cloud LLM provider from a single, self-hosted interface. It ships with multi-user support, a built-in vector database, and document ingestion out of the box, so a private ChatGPT-style experience can be running in minutes with no complicated setup.

AnythingLLM Docker Compose example

Self-host AnythingLLM on your own server, homelab, or VPS starting from this Docker Compose example. It runs AnythingLLM in Docker containers using the official mintplexlabs/anythingllm:latest image, with persistent volumes and automatic restarts preconfigured. Review the environment variables and adjust them to your setup, save the file as compose.yml (or docker-compose.yml), and start the stack with docker compose up -d.

services:
  anythingllm:
    image: mintplexlabs/anythingllm:latest
    # Required by the built-in document collector: it runs a headless Chromium
    # to scrape/parse web pages and files, which needs the SYS_ADMIN capability.
    cap_add:
      - SYS_ADMIN
    volumes:
      - storage:/app/server/storage
    restart: unless-stopped
    environment:
      # Directory inside the container that holds all persistent data:
      # the SQLite database, the LanceDB vector store, and uploaded documents.
      STORAGE_DIR: "/app/server/storage"
      # Secret used to sign multi-user session tokens. Set this to a long,
      # random string when enabling password protection / multi-user mode.
      JWT_SECRET: "changeme"

volumes:
  storage:

Values set to changeme are required — replace them with your own values before starting AnythingLLM.

Prefer a managed setup? WinterFlow installs, configures, and updates AnythingLLM for you using this same Docker Compose configuration.