Finally! Thanks LinkedIn for that CROP Finishing my first agent with eternal memory (until proven otherwise): - Connects nodes using Graph Memory - Manages infinite users determined by ID, Email, Conversations by Session_ID, and messages by Timestamp (YYYY:MM:DD:HH:MM:SS). - If the user changes something in their history, for example, they used to like ice cream and say, "I don't like ice cream," it updates in the Graph. - GPT-4o Mini model - Long-term memory is stored in a MongoDB Cluster (512MB) - The VectorDB for quick search with semantic search is Pinecone, on AWS. - Integrated with LangChain to enhance the agent with different tools - The most basic front end possible with Next - Next and last: Integrate with LangFlow to multiply agents, memories, and tools. Only the information entered by me has a weight of 70kb. Let's think about what we do with 5gb. 50. 1tb. Agent with potential for customer experience or support. I would like to know if they have potential for any branch of medicine. 🤔 #aiagents #llm #aiautomation #graphmemory #aiforbusiness