Weaviate Makes Engram Generally Available, Giving AI Agents Production-Grade Memory

Weaviate, a leading company in the open-source AI-native database space, has officially announced the general availability of Engram, its innovative managed memory and context service tailored for agentic applications. This new offering is now accessible through Weaviate Cloud, which also includes a free tier, making it easier for developers and organizations to integrate advanced memory capabilities into their AI agents. The introduction of Engram addresses a critical challenge faced by organizations as they scale AI agents into production environments: the inability of agents to reliably retain information across different sessions, learn effectively from user feedback, or seamlessly share contextual data when tasks involve multiple agents. Current workarounds often involve cumbersome methods like replaying entire conversation histories or manually curating files, which are not scalable for real-world complexities. Engram fundamentally redefines how memory is handled for AI agents by treating it as a first-class infrastructure component, on par with traditional storage and retrieval mechanisms. Instead of merely accumulating an ever-growing context, Engram employs asynchronous pipelines to actively maintain memory. These pipelines extract pertinent facts from raw agent interactions, reconcile them against existing knowledge, and manage deduplication and updates to ensure that memories are clean, structured, and durable. This "fire-and-forget" approach allows applications to hand off raw events and continue processing while memory is built robustly in the background. A core feature of Engram is its ability to serve these memories through Weaviate's hybrid search capabilities. This means that teams can leverage a memory layer powered by their trusted retrieval stack, ensuring efficient and context-aware data access. Furthermore, Engram emphasizes scoping, isolating memories per project, user, and property. This granular control ensures that the right context is delivered only to the appropriate caller, enhancing security and relevance. Common use cases for Engram include enabling personalization in AI applications, facilitating continual learning for agents, and managing shared states in multi-agent systems. Weaviate provides ready-made templates for these scenarios, with options for customization to meet specific project requirements. Bob van Luijt, CEO and co-founder of Weaviate, highlighted the significance of memory, stating, "Memory is the difference between an agent that answers a question and an agent that gets better at its job." He further emphasized that Engram provides a production-grade memory layer without the overhead of building and operating one from scratch. The service is now fully available, with a free tier offering 1,000 pipeline runs per month, and paid plans starting at $45 per month. Weaviate's platform, which includes the Weaviate Database, Query Agent, and Engram, continues to support the development of scalable AI applications, with its open-source database seeing over 150 million downloads monthly.