major7apps/pensyve
📦 Open Source Projectmajor7apps
A high-performance, universal memory runtime built in Rust to provide persistent, context-aware storage for AI agents.
Pensyve provides a robust infrastructure for AI agent memory, moving beyond simple RAG implementations. At its core, it leverages Rust for high-performance state management and low-latency retrieval. The system utilizes a hybrid approach to memory, combining vector embeddings for semantic similarity with knowledge graphs to map relationships between entities and concepts. This dual-structure allows agents to perform both fuzzy semantic lookups and precise relational queries. Key features include an integrated spaced-repetition engine that prioritizes relevant information over time, ensuring that the most critical context remains accessible while pruning noise. Designed as a universal runtime, it supports various LLM backends and integrates with modern agent frameworks, making it a versatile tool for developers building sophisticated, stateful AI applications that require deep, evolving context.
💡Highlights
- ├─Hybrid vector and graph storage
- ├─Spaced-repetition memory logic
- └─High-performance Rust runtime
🎯For
- ├─AI Agent Developers
- └─Backend Engineers