Show HN: Memoripy – AI Memory Layer with Short- & Long-Term Memory, Clustering
github.comHey HN! I built Memoripy, a memory layer for AI that adds short-term, long-term, and semantic memory capabilities to enhance LLM applications. It helps AI systems retain and prioritize past interactions, adapt over time, and respond with greater context and personalization. Memoripy uses semantic clustering to retrieve relevant memories, along with adaptive memory decay and reinforcement, so interactions stay fresh and context-aware. It’s designed for easy integration with OpenAI, Ollama, and other platforms—giving your AI applications dynamic memory management with minimal setup.
Would love to hear your thoughts and feedback!
Congrats on launching! Looks pretty neat.
I like the fact that BaseStorage is abstract. `sqlite-vec` had been pretty good at what it does, in case you're looking for a good built-in vector search.
Thank you! Will look into that
Cool project, looks neat!
I'm curious how the adaptive memory decay is implemented and how is it different from mem0?
Thanks for the kind words! The adaptive memory decay in Memoripy works by dynamically adjusting the relevance of stored memories based on time, frequency of access, and contextual importance. Over time, less frequently accessed memories decay, ensuring the system stays focused on what’s most relevant without overwhelming the AI with outdated information.
Unlike Mem0, which uses a hybrid database system and scoring layers to prioritize data, Memoripy applies decay and reinforcement directly within its memory layer, integrating semantic clustering to group related memories for more context-aware retrieval. This makes the approach more lightweight and adaptive, especially for applications that don’t require complex multi-database setups.
Hope that clears it up! Let me know if you'd like more details.