Blog
Mindcache is being built in the open through architecture notes, runtime specs, and product thinking. This page will grow into the company's research log. For now, it begins with the memory runtime v0 design document.
Why not all memory should be structured at import time, why flat capture matters for backfill, and why human-facing topic structure should emerge slowly through digestion.
Read the documentWhy Mindcache moved away from graph-first memory: sources as materials, fact/event/plan atoms, daily writeups, and topic-tree-first retrieval.
Read the documentThe first public design document for Mindcache's local-first memory runtime: graph as core, tree as view, budgeted retrieval, and nightly digest.
Read the document