Repolode
Submit
langflow-ai

langflow-ai/openrag

A self-hosted RAG platform that lets you upload documents and chat with them using LLMs, built on Langflow, Docling, and OpenSearch.

View on GitHubWebsite
Stars
4,674
Forks
518
Open issues
347
Last commit
14 hours ago

What is it?

OpenRAG is a single-package Retrieval-Augmented Generation platform for searching documents and holding AI-powered conversations over them. Users upload, process, and query files through a chat interface backed by large language models and semantic search.

The system uses Langflow for document ingestion, retrieval workflows, and agent-style orchestration, Docling for parsing documents, and OpenSearch as the search and storage layer. The backend is built with FastAPI and the frontend with Next.js. The project can be installed as a Python package with the uv tool or deployed as self-managed services with Docker or Podman.

OpenRAG also exposes a REST API through official Python and TypeScript SDKs, and a built-in MCP server over streamable HTTP at the /mcp endpoint for connecting clients such as Cursor and Claude Desktop.

Key features

  • 01Chat interface for asking questions over uploaded documents using LLMs
  • 02Document ingestion and parsing with Docling
  • 03Visual workflow builder powered by Langflow
  • 04OpenSearch-backed semantic search at scale
  • 05Official Python and TypeScript SDKs
  • 06Built-in MCP server for connecting AI assistants

Use cases

  • Building a private question-answering assistant over internal documents
  • Searching a team's file collection by meaning rather than keywords
  • Connecting an AI coding or chat assistant to a personal or company knowledge base
  • Prototyping RAG pipelines with visual Langflow workflows

Quick start

  1. 1

    Install OpenRAG by following the Quickstart guide in the docs

  2. 2

    Launch OpenRAG

    uv run openrag
  3. 3

    Add files or folders as knowledge

  4. 4

    Start chatting with your knowledge in the web interface

Things to know

  • It is a multi-service stack built on OpenSearch, Langflow, and Docling, so self-hosting takes real setup effort.
  • The standalone openrag-mcp PyPI package is deprecated, so MCP clients must connect to the /mcp endpoint instead.

Topics

#rag#llm#document-search#vector-search#opensearch#langflow#python#fastapi

FAQ

What is openrag?
A self-hosted RAG platform that lets you upload documents and chat with them using LLMs, built on Langflow, Docling, and OpenSearch. OpenRAG is a single-package Retrieval-Augmented Generation platform for searching documents and holding AI-powered conversations over them. Users upload, process, and query files through a chat interface backed by large language models and semantic search.
Is openrag free to use? What license does it have?
openrag is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is openrag still maintained?
Actively maintained. The last commit to openrag was on Nov 6, 2026.
What can I use openrag for?
Building a private question-answering assistant over internal documents Searching a team's file collection by meaning rather than keywords Connecting an AI coding or chat assistant to a personal or company knowledge base Prototyping RAG pipelines with visual Langflow workflows
How do I install openrag?
Install OpenRAG by following the Quickstart guide in the docs Launch OpenRAG: uv run openrag Add files or folders as knowledge Start chatting with your knowledge in the web interface
Where can I get openrag?
The source code and installation instructions are on GitHub: https://github.com/langflow-ai/openrag

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects