OpenRAG is a comprehensive, single-package Retrieval-Augmented Generation platform built by the Langflow team. It combines document ingestion, search and chat into one deployable system, so you can go from a pile of documents to a working question-answering application without assembling the stack yourself.
OpenRAG is built on three proven open-source components: Langflow provides the visual, agentic workflow layer, Docling handles document parsing and conversion, and OpenSearch powers indexing and retrieval. On top of that foundation it offers document ingestion pipelines, a chat interface for asking questions over your corpus, and agentic workflows with capabilities such as reranking and multi-agent coordination. Because it is based on Langflow, flows can be built and adjusted in a drag-and-drop visual builder, while developers get Python and TypeScript SDKs for programmatic control. A FastAPI backend and Next.js frontend round out the package, and a built-in MCP server lets external agents use your RAG pipeline as a tool. Everything is designed to be self-hosted as one package, which suits organizations that cannot send their documents to third-party hosted retrieval services. Deployment is aimed at a single server or container host, keeping operations simple for small teams. The project is Apache-2.0 licensed, has earned more than 4,600 stars on GitHub, and was under active development in 2026.
Free. OpenRAG is open source under the Apache-2.0 license and is self-hosted; you bring your own infrastructure and model API keys.
Developers and teams who want a complete, self-hosted RAG platform for document question answering, and Langflow users who want retrieval, ingestion and chat packaged together instead of wiring the components up manually.