OneRAG and raggo
These are competitors offering production-ready RAG frameworks in different languages (Python/FastAPI vs Go), each providing their own abstractions for swapping vector databases and LLM providers rather than interoperating together.
About OneRAG
notadev-iamaura/OneRAG
Production-ready RAG Framework (Python/FastAPI). 1-line config swaps: 6 Vector DBs (Weaviate, Pinecone, Qdrant, ChromaDB, pgvector, MongoDB), 5 LLMs (Gemini, OpenAI, Claude, Ollama, OpenRouter). OpenAI-compatible API. 2100+ tests.
Supports hybrid search (dense + BM25), GraphRAG for knowledge graph reasoning, and pluggable rerankers (6 options including Jina and Cohere) through a modular pipeline architecture. Includes built-in PII detection/masking, semantic/Redis caching layers, and query routing that classifies requests before retrieval. Designed for gradual complexity—start with basic vector search and layer in advanced features like agents and tool execution without refactoring the codebase.
About raggo
teilomillet/raggo
A lightweight, production-ready RAG (Retrieval Augmented Generation) library in Go.
Supports pluggable vector stores (Milvus) and embedding providers (OpenAI), with modular components for document loading, chunking, and embedding that can be composed into different RAG pipelines. Offers multiple implementations—SimpleRAG for basic Q&A, ContextualRAG for semantic understanding with automatic context generation, and MemoryContext for chat applications—allowing developers to choose the complexity level needed. Configuration is flexible, supporting environment variables, JSON files, and programmatic setup with features like hybrid search, chunk overlap control, and similarity thresholds for fine-tuned retrieval behavior.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work