A modular multi-modal transactional database for AI and semantic search, replacing MongoDB, Neo4J, and Elastic with a single ACID solution.
UStore is a modular, multi-modal transactional database that consolidates the functionality of specialized databases like MongoDB, Neo4J, and Elasticsearch into a single, faster ACID-compliant solution. It is built for artificial intelligence and semantic search workloads, supporting blobs, documents, graphs, and vectors with high-performance backends and familiar APIs.
Developers and data engineers working on AI, semantic search, or complex data applications that require unified storage for multiple data modalities (documents, graphs, vectors) with transactional guarantees.
UStore offers a single, high-performance ACID database that replaces multiple specialized systems, reducing complexity and infrastructure costs. Its modular design, familiar APIs (like Pandas/NetworkX), and support for open standards make it both flexible and developer-friendly.
Multi-Modal Database replacing MongoDB, Neo4J, and Elastic with 1 faster ACID solution, with NetworkX and Pandas interfaces, and bindings for C 99, C++ 17, Python 3, Java, GoLang 🗄️
Open-Awesome is built by the community, for the community. Submit a project, suggest an awesome list, or help improve the catalog on GitHub.
UDisk and UCSet engines offer 10x to 30x speed improvements over RocksDB, with kernel-bypass capabilities for extreme throughput, as shown in performance charts.
Supports blobs, documents, graphs, and vectors in a single ACID-transactional system, aiming to replace specialized databases like MongoDB and Neo4J.
Provides Python interfaces mimicking Pandas DataFrames and NetworkX graphs, lowering the learning curve for data scientists and engineers.
Uses Apache Arrow Flight for remote access and RFC standards for JSON operations, promoting interoperability and avoiding vendor lock-in.
Horizontal scaling is listed as upcoming in the roadmap, making current versions unsuitable for distributed workloads without custom sharding.
Drivers for GoLang and Java lack support for documents and graphs, limiting full multi-modal access in those ecosystems.
Requires detailed configuration files and engine-specific optimizations, which can be challenging for teams without dedicated database administration.
UStore is an open-source alternative to the following products:
Neo4j is a graph database management system that uses graph structures with nodes, edges, and properties to represent and store data.
Elasticsearch is a distributed, RESTful search and analytics engine capable of addressing a growing number of use cases, built on Apache Lucene.
Pinecone is a vector database service designed for machine learning applications, enabling efficient storage and retrieval of high-dimensional vector embeddings.