An operating system for long-lived agent memory — hierarchical, explainable, and self-managing. Replace passive vector dumps with human-like memory layers.
Most agent memory today is RAG in a trench coat. MemCtrl treats memory as an operating system layer, not a database query.
| Capability | Vector RAG | MemCtrl |
|---|---|---|
| Retrieval logic | Cosine similarity (black box) | Hierarchical tree traversal with trace |
| Explainability | "Score: 0.87" | root -> project -> backend -> fastapi |
| Lifespan control | Manual cleanup | Rule-driven expiry + never-forget lists |
| Knowledge consolidation | None | Automatic session -> project merging |
| Audit trail | None | Complete trigger log |
| Privacy | Cloud embeddings | Local SQLite only |
MemCtrl implements a cognitive memory pipeline inspired by how humans actually remember.
Chat, code, events, or files enter the system
Secrets and PII are redacted before storage
LLM extracts facts with confidence scoring
Project (forever), Session (7 days), User (90 days)
Auto-summarize and consolidate experience
LLM reasons over structure, not vectors
We measure what actually matters for agent memory systems.
The building blocks of agent cognition.
Project, session, and user layers with different lifespans. Like human semantic, working, and episodic memory.
Automatically refine and summarize memories. Session notes consolidate into permanent project knowledge.
Distill verbose conversations into dense, actionable facts without losing meaning.
Every answer includes the exact reasoning path taken through the memory tree. Fully explainable.
Context-aware retrieval that considers recency, relevance, and confidence together.
Secrets and PII are redacted before storage. Never-forget lists protect sensitive memories.
Join developers building the future of AI cognition.