# Memory

> Browse 5 items tagged with Memory.

## Items

- [Zep](https://vectordb.ever.works/de/items/zep) — Context engineering and agent memory platform for AI agents with sub-200ms latency. Zep uses a temporal knowledge graph architecture to deliver relationship-aware context from chat history, business data, documents, and app events.
- [ospipe](https://vectordb.ever.works/de/items/ospipe) — RuVector-enhanced personal AI memory for Screenpipe, replacing SQLite with semantic vector search, knowledge graphs, and attention reranking.
- [Compression Ratio Optimization](https://vectordb.ever.works/de/items/compression-ratio-optimization) — Techniques for optimizing the trade-off between memory usage and accuracy in vector quantization, achieving 5-40x compression in systems like Mastra's Observational Memory.
- [Letta](https://vectordb.ever.works/de/items/letta) — Platform for building stateful AI agents with advanced memory that can learn and self-improve over time. Uses OS-inspired approach with main context as RAM and external storage as disk.
- [Observer-Reflector Architecture](https://vectordb.ever.works/de/items/observer-reflector-architecture) — Memory system architecture used in Mastra's Observational Memory with two background agents that compress and garbage collect conversation history achieving 5-40x compression.

---

_Canonical page: https://vectordb.ever.works/de/tags/memory_