Build memory that can show its work.
CoreCortex is a control plane and contract for consequential retrieval: configure the corpus, preserve the evidence, choose storage deliberately, and measure every change.
Build a corpus
Start with an authored configuration, validate it, and inspect the query plan.
Getting started →02Understand the model
Learn how records, fragments, relations, provenance, and versions fit together.
Core concepts →03Integrate retrieval
Use explicit retrieval plans and return ranked evidence to your application.
Querying →What CoreCortex is—and is not
CoreCortex is not another chat interface or a database with a new label. It is the versioned layer that connects source records to derived evidence, storage indexes, retrieval policy, and evaluation.
Core can be run and extended by your team. CoreCortex Enterprise operates source ingestion and continuous strategy experiments for teams that do not want to build that machinery.