rush-db/rushdb
RushDB is an open-source database for AI. It records fields, types and values on every write, so agents can read the schema before they query.
What it solves
RushDB eliminates the need for complex schema design. It allows you to store unstructured JSON data while automatically inferring types and mapping nested objects into a searchable graph structure, saving you from manually defining tables or relationships.
How it works
When you ingest JSON, RushDB treats every nested object as a distinct record linked to its parent. It stores this as a Labeled Meta-Property Graph (LMPG), which keeps track of properties and relationships automatically. You can then interact with this data using a wide variety of interfaces, including SQL, Cypher, Gremlin, Mongo-style documents, and a native JSON query language, all within a single Rust binary.
Who it’s for
It is designed for developers who want a flexible, multi-model database that handles both structured records and complex relationships without the overhead of managing rigid schemas or multiple specialized database systems.
Highlights
- Multi-model support: Query the same data using SQL, Cypher, Gremlin, or native JSON.
- Built-in search: Includes vector indexes for AI applications, full-text search, and hybrid search capabilities.
- Operational simplicity: Runs as a single binary with no external dependencies required.
- Enterprise-ready features: Supports multi-statement transactions, point-in-time recovery, and streaming replication.
- Protocol versatility: Communicates via HTTP, the PostgreSQL wire protocol, gRPC, and Bolt.
Related
- Project
- Project
- Project
- Dispatch
- Project