Show HN: I made an open-source memory layer for agents (agentrecall.cloud)

A new open-source memory layer for AI agents has been announced, leveraging Neo4j to create a relationship graph that connects memories rather than storing them in isolation. This innovative system allows AI agents to traverse memories, uncovering hidden connections and interactions, which can enhance decision-making and operational efficiency. For instance, with queries structured to explore specific topics, agents can dynamically access contextually relevant information. This memory layer significantly impacts the AI/ML community by enabling agents to automatically extract entities, categorize memories, and perform semantic searches using vector embeddings, moving beyond simple keyword matching. The system features a cloud-based API that supports multiple agents, allowing each to maintain an isolated memory while also enabling cross-agent collaboration. Additionally, users can deploy the layer in either local or cloud modes, retaining flexibility based on their infrastructure needs. This architectural design and functionality promise to enhance the collaborative capabilities of AI systems, making them more adaptable and efficient in various applications.