nmemo aiMemory infrastructure

The memory layer for AI apps

Neutral memory backend behind three calls: store, recall, and govern what your users tell your AI. Portable identity across apps, supersede chains with provenance, consolidation sleep cycles so context doesn't rot into noise.

nmemo.add()Ingest & extract

Conversations in facts out, async, deduped

nmemo.recall()Hybrid retrieval

Relevance × recency × importance, token-capped

nmemo.forget()Governed erase

Per-memory delete with provenance & audit

Not a vector-store bolt-on. Not an agent OS. Per-fact audit, PII redaction, per-tenant token budgets: memory that follows the user and proves what it knows.

Governed, self-maintaining memory across every app you build.