Today's RAG & Vector Databases: Fastest-Growing Projects — September 03, 2026
Today's the RAG & Vector Databases space, self-hosted knowledge platforms and vector databases continue to gain traction, with several projects showing significant growth and engagement on GitHub. The most notable trend involves Rust-based tools that offer robust document management features, while Go-based libraries provide high-performance alternatives for vector search and storage.
deeplethe/utopia: This project offers a self-hosted knowledge platform where Retrieval-Augmented Generation (RAG) is applied over documents stored in a knowledge graph. The graph retains information about when each fact was true, providing a dynamic view of evolving data. With a growth score of 62.93 and 2,228 stars, Utopia's popularity can be attributed to its unique approach to managing temporal facts within a scalable Rust-based system.
rostamlabs/rostam: Rostam is an open-source vector database that also functions as a sub-microsecond key-value store, offering various indexing methods and hybrid search capabilities. It supports both dense and sparse searches alongside BM25 relevance ranking, making it versatile for different use cases. With 100 commits in the last month and a growth score of 13.21, Rostam's steady development and robust feature set are driving its appeal among developers seeking high-performance vector search solutions.
makralabs/makra: Makra serves as an intermediary layer between web data and AI agents, providing real-time structured data through vector search capabilities. This project, with 28 stars and a growth score of 7.00, is growing due to its innovative approach in bridging the gap between unstructured web content and AI applications that require precise data inputs.
Shuo-Liang-0111/RA-Bench: RA-Bench focuses on benchmarking techniques for detecting AI-generated videos in real-world crisis scenarios, anchoring events to their actual occurrences. Despite a lower growth score of 2.98, the project's relevance to current issues surrounding deepfake detection and misinformation makes it an important tool for researchers and security professionals.
While these projects showcase diverse applications within RAG and vector databases, they collectively highlight the growing interest in self-hosted solutions that offer robust data management and retrieval capabilities.
deeplethe/utopia: This project offers a self-hosted knowledge platform where Retrieval-Augmented Generation (RAG) is applied over documents stored in a knowledge graph. The graph retains information about when each fact was true, providing a dynamic view of evolving data. With a growth score of 62.93 and 2,228 stars, Utopia's popularity can be attributed to its unique approach to managing temporal facts within a scalable Rust-based system.
rostamlabs/rostam: Rostam is an open-source vector database that also functions as a sub-microsecond key-value store, offering various indexing methods and hybrid search capabilities. It supports both dense and sparse searches alongside BM25 relevance ranking, making it versatile for different use cases. With 100 commits in the last month and a growth score of 13.21, Rostam's steady development and robust feature set are driving its appeal among developers seeking high-performance vector search solutions.
makralabs/makra: Makra serves as an intermediary layer between web data and AI agents, providing real-time structured data through vector search capabilities. This project, with 28 stars and a growth score of 7.00, is growing due to its innovative approach in bridging the gap between unstructured web content and AI applications that require precise data inputs.
Shuo-Liang-0111/RA-Bench: RA-Bench focuses on benchmarking techniques for detecting AI-generated videos in real-world crisis scenarios, anchoring events to their actual occurrences. Despite a lower growth score of 2.98, the project's relevance to current issues surrounding deepfake detection and misinformation makes it an important tool for researchers and security professionals.
While these projects showcase diverse applications within RAG and vector databases, they collectively highlight the growing interest in self-hosted solutions that offer robust data management and retrieval capabilities.