For a hybrid retrieval engine for RAG, the first results are bragai/brag-langchain, hkuds/lightrag and tobi/qmd (qmd is a local hybrid search engine that combines keyword and vector retrieval with reranking, exposed as an MCP server or HTTP API, and is purpose-built for RAG pipelines — exactly the kind of self-hostable, hybrid retrieval system this search targets). elastic/elasticsearch and lancedb/lancedb round out the shortlist. Compare the match explanations and check the project documentation against your requirements.
Open-source frameworks and libraries implementing combined keyword and vector retrieval for advanced retrieval-augmented generation systems.
bRAG-langchain is a framework for building retrieval augmented generation pipelines using LangChain to connect documents with language models. It functions as a vector store orchestrator that manages document indexing and retrieval strategies to improve context accuracy. The system implements an advanced retrieval pipeline featuring a semantic query router that directs natural language inputs to specific data sources or prompts. It includes a metadata filtering engine that translates natural language queries into structured schemas to narrow search results. The project covers hybrid search o
bRAG-langchain is a LangChain-based framework for building RAG pipelines that includes hybrid search with semantic routing, metadata filtering, and retrieval fusion, making it a fitting choice for integrating sparse and dense retrieval into your RAG setup, though you may need to wrap it for a standalone query API.
LightRAG is a graph-based retrieval framework designed to build retrieval-augmented generation pipelines. It structures unstructured text into knowledge graphs, enabling multi-hop reasoning and complex query synthesis across large document collections. By integrating dense vector embeddings with structured knowledge graphs, the system facilitates both similarity-based and relationship-aware information retrieval. The framework distinguishes itself through a dual-level retrieval strategy that combines low-level keyword matching with high-level semantic graph traversal to capture both specific
LightRAG is a graph-based retrieval framework that combines low-level keyword matching (sparse retrieval) with dense vector embeddings and structured knowledge graphs in a dual-level retrieval strategy, purpose-built for RAG pipelines, self-hostable, and supporting both similarity-based and relationship-aware search—exactly the kind of hybrid retrieval system this query targets.
qmd is a local semantic search engine and RAG knowledge base indexer that functions as a Model Context Protocol server. It converts local documents, markdown files, and codebases into a searchable database to provide retrieval augmented generation capabilities for AI agents. The system exposes its search and retrieval tools via stdio or HTTP. It utilizes local model files for embeddings and reranking, supporting query expansion across multiple languages. The project employs abstract syntax tree based chunking to split source code at function and class boundaries. It implements hybrid vector-
qmd is a local hybrid search engine that combines keyword and vector retrieval with reranking, exposed as an MCP server or HTTP API, and is purpose-built for RAG pipelines — exactly the kind of self-hostable, hybrid retrieval system this search targets.
Elasticsearch is a distributed search engine and document store designed for the high-performance indexing and retrieval of massive volumes of unstructured data. It functions as a centralized analytics platform, providing a schema-flexible architecture that organizes information into searchable indices while maintaining global cluster state through a distributed consensus mechanism. The platform distinguishes itself through its integrated approach to observability, security, and advanced analytics. It combines full-text, vector, and hybrid search capabilities with machine learning-driven insi
Elasticsearch is a distributed search engine that natively combines full-text BM25 and dense vector retrieval with hybrid search and reranking, offers a query API, is self-hostable, and is widely used as the retrieval component in RAG pipelines.
LanceDB is a vector database and columnar data store designed to function as a versioned dataset manager and vector search engine. It serves as a high-performance backend for indexing and retrieving high-dimensional embeddings, providing the foundation for machine learning data pipelines. The system distinguishes itself through a combination of cloud-native object storage and immutable version tracking, allowing for data time-travel and reproducible AI experiments. It integrates hybrid search capabilities, merging dense vector similarity with BM25 full-text search and SQL-like scalar filters
LanceDB is a self-hostable vector database that natively combines dense vector similarity with BM25 full-text search and scalar filtering, making it a ready-to-use hybrid search engine well-suited for integration into RAG pipelines via its API and configurable retrieval fusion.
This project is a retrieval augmented generation framework designed to build pipelines that connect unstructured data and knowledge graphs with large language models. It functions as a vector database orchestrator for indexing text and multimodal content, as well as a system for translating natural language queries into structured database commands. The framework integrates a hybrid retrieval engine that combines dense vector search with sparse keyword matching to increase the precision of retrieved contexts. It further enhances reasoning and relationship mapping through a graph-augmented ret
This repository is a RAG framework that includes a hybrid retrieval engine combining dense vector search and sparse keyword matching, directly supporting the desired sparse-dense fusion for RAG pipelines.
Vespa is a distributed search engine, vector database, and machine learning ranking engine. It serves as an AI search platform designed to handle large-scale document indexing and complex query processing across a cluster of nodes, combining keyword retrieval with high-dimensional embedding storage for semantic similarity search. The platform distinguishes itself by integrating machine learning models directly into the search pipeline to perform real-time inference and ranking. It converts these models into ranking expressions to score and order results based on relevance, while providing a s
Vespa is a distributed search engine and vector database that natively blends BM25 keyword retrieval with dense vector search and ML-powered reranking, making it an ideal platform for building self-hosted hybrid RAG pipelines with a full query API.
Meilisearch is a Rust-based search engine providing typo-tolerant full-text and vector-based semantic search with real-time conversational capabilities.
Meilisearch is a full-featured hybrid search engine combining full-text (BM25-style) and vector semantic search with a REST API for self-hosted deployment, making it directly usable as the retrieval component in RAG pipelines with configurable hybrid ranking.
localGPT is a private AI knowledge base and retrieval-augmented generation application. It provides a local document indexer, a hybrid search engine, and an inference interface to enable chatting with private documents and managing a self-hosted information repository without sending data to external servers. The system distinguishes itself through a dual-pass verification pipeline that ensures generated answers are grounded in retrieved sources, accompanied by explicit source attribution. It employs a hybrid retrieval approach combining semantic vector search with keyword matching and rerank
localGPT is a self-hostable hybrid search engine that combines dense vector search and keyword matching with reranking, purpose-built for RAG pipelines — directly matching the search for a hybrid retrieval system with local deployment, API access, and source-attributed answers.
Weaviate is an AI-native vector database designed to store and index high-dimensional vector embeddings alongside traditional data objects. It serves as a backend infrastructure for retrieval-augmented generation, enabling applications to ground language model responses in private, context-aware data. The platform distinguishes itself by combining vector similarity search with traditional keyword filtering through a hybrid storage architecture. It integrates directly with external machine learning models to automate the generation of embeddings and perform complex inference tasks during inges
Weaviate is an open-source vector database with native hybrid search that fuses BM25 keyword and dense vector retrieval, offers configurable weighting, and is purpose-built as a backend for RAG pipelines with a full query API and self-hostable deployment.
FastGPT is a comprehensive platform for building, deploying, and managing context-aware artificial intelligence applications. It provides a unified environment that integrates custom data sources with language models, utilizing a retrieval-augmented generation engine to ground responses in accurate, domain-specific information. The system is designed for enterprise-scale use, featuring multi-tenant architecture, administrative controls, and secure authentication protocols including OAuth 2.0 and custom single sign-on integration. The platform distinguishes itself through a visual, node-based
FastGPT is a full-stack RAG platform that includes hybrid search (combining sparse and dense retrieval), reranking, and fusion, with an API for querying and self-hostable deployment, fitting your need for a hybrid search system integrated into RAG pipelines—though it offers a broader application-building environment rather than a pure search engine.
Airweave is a unified AI knowledge base platform that syncs data from external APIs into a searchable layer for retrieval-augmented generation. It provides a pre-built data connector library and a framework for building custom connectors, enabling the extraction, transformation, and synchronization of structured and unstructured data from SaaS applications. The platform includes a hybrid vector retrieval system that combines semantic, neural, and keyword search strategies to deliver grounded context for AI agents. The platform distinguishes itself through an agentic search engine that iterati
Airweave is a hybrid retrieval platform for RAG that syncs external data and combines keyword and semantic search, so it squarely fits the need for a hybrid search engine designed around RAG pipelines.
Chroma is a specialized vector database designed to index and retrieve high-dimensional data representations for semantic similarity search. It functions as a comprehensive platform for information retrieval, enabling the storage and management of unstructured documents alongside structured metadata. By mapping data into numerical representations, the system facilitates rapid similarity lookups across large datasets. The platform distinguishes itself through a hybrid search infrastructure that combines dense vector embeddings with sparse keyword and regular expression matching to balance sema
Chroma is a vector database that natively supports hybrid dense-sparse search (embedding similarity plus keyword/regex matching) and is designed for RAG pipelines, making it a fit for your intent, though its sparse retrieval is not BM25-based and it does not include a dedicated reranking step.
Typesense is a distributed search engine designed to provide sub-millisecond query latency across massive datasets. It functions as both a high-performance indexing and retrieval engine and a comprehensive search experience platform, offering built-in typo tolerance and tools for managing relevance through synonym configuration, result curation, and complex filtering. The platform distinguishes itself by utilizing in-memory indexing to maintain high-throughput data retrieval and integrating vector database capabilities to support semantic similarity searches. It ensures data consistency and h
Typesense is a distributed search engine that supports both full-text keyword retrieval and vector-based semantic search, making it usable as a retrieval backend in RAG pipelines, but it does not explicitly provide built-in hybrid fusion or RAG-specific integration workflows.
| Repository | Stars | Language | License | Last push |
|---|---|---|---|---|
| bragai/brag-langchain | 4.2K | Jupyter Notebook | NOASSERTION | |
| hkuds/lightrag | 40.6K | Python | MIT | |
| tobi/qmd |
| 30.9K |
| TypeScript |
| MIT |
| elastic/elasticsearch | 78.5K | Java | NOASSERTION |
| lancedb/lancedb | 9K | HTML | apache-2.0 |
| datawhalechina/all-in-rag | 4K | Python | — |
| vespa-engine/vespa | 7K | Java | Apache-2.0 |
| meilisearch/meilisearch | 59.8K | Rust | NOASSERTION |
| promtengineer/localgpt | 22.2K | Python | MIT |
| weaviate/weaviate | 16.9K | Go | NOASSERTION |