Supermemory
The company behind company-brain and the memory backend it runs on: a hosted "memory and context engine" (supermemory.ai) pitched as the context stack for AI agents — memory, RAG, user profiles, connectors and file processing behind one API, with a local/offline mode alongside the hosted service. Discovered while ingesting company-brain; not otherwise evidenced in this corpus.
Its docs claim 95% Recall@15 on the LongMemEval benchmark at roughly 720 tokens of added context (a claimed 99.4% reduction) — a maker's own number, unverified here.
Connections
- publisher of company-brain (Apache-2.0, formerly a paid product built on this same memory layer)