Launch · AI & Deep Tech

CortexDB launches AI memory layer

By Startup Enthusiast · 10 April 2026

What they do

CortexDB provides long-term memory for AI systems, preserving data in an immutable event log

The company launched as a memory layer, using a temporal knowledge graph with causality and provenance

It scales horizontally with consistent hashing and leaderless replication, matching single-node to cluster performance

CortexDB launched as a long-term memory layer for AI systems.
It preserves data in an immutable event log, never overwriting or summarizing.
The system builds a temporal knowledge graph with causality and provenance from the event stream.
It retrieves context using a hybrid approach: vector search, full-text matching, graph traversal, and adaptive ranking.
CortexDB scales horizontally with consistent hashing, partition-aware placement, and leaderless replication.
A single-node deployment runs the same code as a hundred-node cluster, with automatic rebalancing when nodes are added.

CortexDB addresses AI agents' statelessness and memory loss.
It claims a significant performance gap over conventional systems in benchmarks.
The founder's background in distributed systems adds credibility to the technical claims.

CortexDB provides a memory layer for AI systems, storing data in an immutable event log.
It builds a temporal knowledge graph with causality and provenance from events.
It retrieves context using hybrid search: vector, full-text, graph traversal, and adaptive ranking.
It scales horizontally using consistent hashing and leaderless replication.
A single-node deployment runs the same code as a cluster, with automatic rebalancing.