Today's RAG & Vector Databases: Fastest-Growing Projects — September 01, 2026
Today's the RAG & Vector Databases space, self-hosted knowledge platforms and high-performance vector databases continue to attract significant attention as developers seek more robust solutions for managing large-scale data repositories. The growth of Rust-based projects like Utopia highlights a growing interest in leveraging advanced database technologies for efficient document retrieval and management.
deeplethe/utopia
Utopia is a self-hosted knowledge platform that enables users to retrieve information from their documents using Retrieval-Augmented Generation (RAG) techniques over a knowledge graph. It remembers when each fact was true, enhancing the accuracy of historical data tracking. With a growth score of 31.32 and an impressive 759 stars, Utopia's rapid rise can be attributed to its innovative approach to document management and its Rust + PostgreSQL architecture that simplifies deployment.
rostamlabs/rostam
Rostam is an open-source vector database designed for sub-microsecond key-value operations. It supports a variety of indexing methods including HNSW, IVF, and Vamana, along with hybrid dense+sparse search capabilities and BM25 text search functionality. With 14.20 growth score and 22 stars, Rostam's growing popularity is likely due to its flexibility in being embedded as a library or run standalone while offering robust vector database functionalities.
parqdb-io/parqdb
ParquetDB is an embedded vector database capable of handling billions of records built entirely on Parquet and Arrow formats. This billion-scale capability makes it highly efficient for large datasets, contributing to its growth score of 10.64 and 59 stars. Developers are drawn to its ability to manage extensive data sets with ease, making it a promising tool for high-performance vector storage.
makralabs/makra
Makra serves as a memory layer between web services and AI agents, providing real-time structured data from the open web through vector search technology. With a growth score of 8.67 and 28 stars, Makra's increasing popularity can be attributed to its unique approach in serving dynamic, up-to-date information directly to AI systems.
Shuo-Liang-0111/RA-Bench
RA-Bench is a benchmarking tool designed specifically for detecting AI-generated videos within real-world crisis settings. With a growth score of 3.18 and 97 stars, RA-Bench's utility in verifying the authenticity of video content during critical events has garnered significant attention from researchers and developers focused on media integrity.
TOPDEV99999/ai-ShopMind
AI-ShopMind is an application that uses vector embeddings stored in Endee to match customer queries with semantically similar product descriptions, generating relevant responses. With a growth score of 0.62 and 29 stars, AI-ShopMind's focus on enhancing e-commerce experiences through advanced search capabilities has attracted interest from developers looking to improve user engagement in online shopping platforms.
Overall, Today's trends underscore the ongoing development and refinement of technologies aimed at improving data retrieval efficiency and accuracy across various applications, particularly those involving large datasets and real-time information processing.
deeplethe/utopia
Utopia is a self-hosted knowledge platform that enables users to retrieve information from their documents using Retrieval-Augmented Generation (RAG) techniques over a knowledge graph. It remembers when each fact was true, enhancing the accuracy of historical data tracking. With a growth score of 31.32 and an impressive 759 stars, Utopia's rapid rise can be attributed to its innovative approach to document management and its Rust + PostgreSQL architecture that simplifies deployment.
rostamlabs/rostam
Rostam is an open-source vector database designed for sub-microsecond key-value operations. It supports a variety of indexing methods including HNSW, IVF, and Vamana, along with hybrid dense+sparse search capabilities and BM25 text search functionality. With 14.20 growth score and 22 stars, Rostam's growing popularity is likely due to its flexibility in being embedded as a library or run standalone while offering robust vector database functionalities.
parqdb-io/parqdb
ParquetDB is an embedded vector database capable of handling billions of records built entirely on Parquet and Arrow formats. This billion-scale capability makes it highly efficient for large datasets, contributing to its growth score of 10.64 and 59 stars. Developers are drawn to its ability to manage extensive data sets with ease, making it a promising tool for high-performance vector storage.
makralabs/makra
Makra serves as a memory layer between web services and AI agents, providing real-time structured data from the open web through vector search technology. With a growth score of 8.67 and 28 stars, Makra's increasing popularity can be attributed to its unique approach in serving dynamic, up-to-date information directly to AI systems.
Shuo-Liang-0111/RA-Bench
RA-Bench is a benchmarking tool designed specifically for detecting AI-generated videos within real-world crisis settings. With a growth score of 3.18 and 97 stars, RA-Bench's utility in verifying the authenticity of video content during critical events has garnered significant attention from researchers and developers focused on media integrity.
TOPDEV99999/ai-ShopMind
AI-ShopMind is an application that uses vector embeddings stored in Endee to match customer queries with semantically similar product descriptions, generating relevant responses. With a growth score of 0.62 and 29 stars, AI-ShopMind's focus on enhancing e-commerce experiences through advanced search capabilities has attracted interest from developers looking to improve user engagement in online shopping platforms.
Overall, Today's trends underscore the ongoing development and refinement of technologies aimed at improving data retrieval efficiency and accuracy across various applications, particularly those involving large datasets and real-time information processing.