Discussion: Complementary Memory Approaches — Redis Checkpointing vs AMC Semantic Memory

[eduardg7]

Hi @redis-developer,

Great work on langgraph-redis — the implementation of Redis checkpoint savers with TTL support is solid. I especially like how you handle Azure/Enterprise compatibility with the proxy architecture.

I'm building AMC (Agent Memory Controller), a memory layer for AI agents. We focus on complementary use cases to your Redis implementation:

• Redis LangGraph: Checkpointing for session state, ephemeral caching, conversation memory middleware
• AMC: Long-term semantic memory with vector search, persistent user knowledge base

Potential synergy:

• Use Redis checkpointers for session state (fast, ephemeral)
• Use AMC for persistent user memory and knowledge base (long-term, semantic)
• Both can coexist in the same agent architecture

We're launching a beta next week (Mar 24) and looking for early adopters. Given your expertise in Redis for agent systems, I'd value your thoughts on whether these approaches could work together — especially for developers who need both fast session caching and persistent semantic memory.

Link: https://github.com/eduardg7/faintech-lab

I'd be interested in your thoughts on:

1. Do you see a need for long-term semantic memory alongside Redis checkpointing?
2. Any interest in documenting how to combine both approaches in LangGraph agents?

Would be great to exchange ideas. No pressure if you're busy — just thought it might be an interesting discussion.

Best,
Eduard
Founder, Faintech Solutions