RAGFlow

What is RAGFlow?

RAGFlow is an open-source RAG (Retrieval-Augmented Generation) engine based on deep document understanding. It offers a streamlined RAG workflow that is meticulously designed for businesses of any scale.

Features

• Deep Document Understanding: Advanced document parsing and chunking capabilities
• Template-based Chunking: Intelligent content segmentation
• Grounded Citations: Trustworthy question-answering with verifiable sources
• Multiple Document Formats: Support for PDF, DOCX, PPT, Excel, images, and more
• Compatibility: Works with heterogeneous model ecosystems including Ollama
• Modular Architecture: Core service with optional additional components

Architecture

This template includes:

• RAGFlow Core: Main application service (required)
• MySQL: Database for storing metadata and configurations (required)
• Optional Services: Redis, MinIO, Elasticsearch, OpenSearch, and Infinity (can be deployed separately if needed)

Environment Variables

The following environment variables will be automatically configured:

• ${RAGFLOW_URL}: The URL of the RAGFlow service
• ${MYSQL_CONNECTION_STRING}: MySQL database connection string
• ${REDIS_CONNECTION_STRING}: Redis connection string
• ${MINIO_ENDPOINT}: MinIO object storage endpoint
• ${ELASTICSEARCH_URL}: Elasticsearch search engine URL
• ${OPENSEARCH_URL}: OpenSearch alternative search engine URL
• ${INFINITY_URL}: Infinity vector database URL

Getting Started

1. After deployment, access RAGFlow through the provided URL
2. Create your first knowledge base
3. Upload documents (PDF, DOCX, images, etc.)
4. Start asking questions based on your documents

Default Credentials

• RAGFlow admin interface will be accessible without initial authentication
• Database and storage credentials are automatically generated and configured