DEVOURED
RushDB 2.0: Memory Infrastructure for the Agentic Era
RushDB 2.0 unifies graph storage, native semantic search, and an Ontology API into a single agent memory infrastructure, eliminating the need for separate vector stores.
What: RushDB 2.0 introduces native vector support for semantic search, an Ontology API for authoritative schema discovery by agents, MCP access with OAuth 2.0, pre-written "Skills" for common agent tasks, analytical queries with select + groupBy, and the option to "Bring Your Own Neo4j" instance.
Why it matters: This release indicates a maturation in the infrastructure for AI agents, moving towards integrated, schema-aware memory systems that simplify agent development and reduce hallucinations by providing structured context.
Takeaway: Developers building AI agents might investigate RushDB 2.0 for a unified memory solution that promises to reduce complexity compared to stitching together multiple data stores.
Deep dive
- RushDB 2.0 integrates native semantic search directly into its graph database, managing embedding indexes internally and supporting custom embedding models.
- The new Ontology API provides agents with an authoritative schema, including labels, types, value ranges, and relationship maps, to prevent hallucinations.
- An MCP (Multi-Client Protocol) server with OAuth 2.0 is included, exposing over 35 RushDB tools for discovery, read, write, AI, and analytics operations to agents.
- "Skills" are pre-written knowledge files for AI coding agents, covering topics like agent memory patterns, query building, data modeling, and faceted search.
- Analytical queries now support select + groupBy with functions like sum, avg, min, max, collect, and timeBucket, directly within the query layer.
- Users can connect their own Neo4j instances to RushDB, allowing data to remain within their infrastructure using the LMPG model.
- A new "Knowledge Unit" (KU) pricing model is introduced, reflecting data creation and querying, with a more generous free tier.
Decoder
- MCP (Multi-Client Protocol): A standardized protocol for AI agents to interact with tools and services.
- LMPG (Labeled Property Graph): An extension of the property graph model where nodes and relationships can have explicit type labels, improving schema clarity and query performance.
- Ontology API: An API that provides an authoritative, machine-readable schema of the data, helping AI agents understand available labels, types, and relationships.
Original article
Full article content is not available for inline reading.
Explore all 61 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi