An all-in-one AI application for chatting with documents and running AI agents, designed to run locally by default with multi-user permissioning. It supports desktop, Docker, source, and hosted deployment methods.
Project overview
It combines document chat, AI agents, and multi-user access controls in a single application that runs locally by default without requiring coding skills.
Project type
AI Agent · RAG · Vector Database
Use cases
Chat Assistants · Knowledge Q&A
Deployment
Docker / Docker Compose
License
MIT
Best for
General users who need a private, fully-featured AI application for chatting with documents and running agents locally or in the cloud without compromising security.
Developers or teams needing multi-user permissioning and a Developer API to build custom integrations.
Key capabilities
Ingest documents (PDF, TXT, DOCX) and interact with them via a chat UI with source citations and drag-and-drop uploads.
Automate complex workflows with built-in agents capable of browsing the web and utilizing no-code AI Agent builder flows.
Supports multiple users with per-user access control and permissioning without compromising instance security or privacy.
Automatically routes chats to the best provider and model based on user-defined rules.
Allows the LLM to remember important information about the user or workspace.
Run recurring tasks or prompts on a cron schedule with full agent capabilities.
Enables unlimited tools for models while reducing token usage by up to 80% per query.
Compatible with Model Context Protocol (MCP).
Limitations and risks
Inference may route to external cloud models depending on user configuration, and an external PostHog telemetry collection service is documented.
GPU requirements and minimum hardware specifications are not documented.
Operation may depend on optional paid external model providers or services depending on user configuration.
Getting started
Users can download and run the desktop application, which is provided for Mac, Windows, and Linux and runs locally by default with zero setup friction.
Alternatives and comparisons
A self-hostable AI second brain that chats with local or online LLMs and performs semantic search over personal documents across multiple devices.
A single private AI workspace that unifies chat, visual workflows, long-running task management, and over 150 integrations using your own keys and compute.