How myai organizes knowledge
Folders, typed relationships, and a memory that reasons over how your work connects, so the agent sees a structured graph instead of a flat list.
The richer myai's picture of how your work connects, the better its answers. A wave of changes deepened that picture.
Folders and typed relationships
Organize artifacts into nested folders inside a Dimension, and define your own relationship types between them. Related artifacts render as oriented groups, so you see what feeds into what instead of a flat, repetitive list.
The agent reasons over relationships first
Before answering, myai maps the relationships around the artifacts in play: what feeds what, which templates produce which records, what depends on what. It carries a working memory of the folders, types, and connections it has touched, so reasoning stays coherent across a long conversation, and it can see a couple of hops out from a collaborator or partner rather than treating the graph as ending at the first link.
Model types and instances
Templates are framed as model types: schemas that produce instances. You can assign any artifact, a PDF, a canvas, a document, to a model type after the fact, so structure can emerge as patterns appear instead of being designed up front. Creating from a template now pulls in its connected pieces automatically, so a new instance arrives wired up rather than isolated.
Archive without losing anything
Archiving a template hides its instances everywhere at once, without rewriting them, and unarchiving brings them back exactly as they were, edges intact. Bulk archive and unarchive refresh open views immediately, with no ghost rows left behind.
Reliability: the codification step now stays quiet on roughly half of reflections instead of generating filler, keeping the audit trail meaningful, and engagement records can attach to dimensions, workflows, and templates rather than only to instances.