Mem0

Ever wanted your Agents to know about past conversations and user preferences? By default, large language models (LLMs) are stateless, which means they do not remember past conversations. This can make it difficult to create truly personal and helpful AI applications. Excited to share Integrating Long-Term Memory with Gemini 2.5 a new guide on how to add long-term memory to your Gemini 2.5 chatbot using the Gemini API and Mem0: - 🛠️ Complete walkthrough of installing `google-genai` SDK and `mem0ai` with configuration examples - 💾 Configure LLM (Gemini 2.5 Flash) and embedding model (text-embedding-004) with vector stores - 📚 Practical examples of adding conversation history and dummy conversations to build user profiles - 🔍 How to retrieve relevant memories using semantic search and integrate them into responses - 🤖 Full implementation of a memory-enabled chatbot that remembers context across sessions Guide: https://lnkd.in/ePMMM-9C

🚀 Just submitted our project for the 100 Agents Hackathon — meet: 🎯 IntelliSupply — the AI-powered command center for supply chain resilience. 🧠 Built as a multi-agent system, IntelliSupply empowers organizations to simulate, forecast, and mitigate disruptions — before they strike. Think of it as a digital twin + autonomous crisis strategist + real-time intelligence, all in one. 🌍 The impact? Supply chain disruptions cost global businesses over \$1.9 trillion annually. IntelliSupply helps organizations proactively identify threats, run simulations, and deploy AI-driven mitigation — reducing losses, improving recovery time, and enhancing competitive advantage. With IntelliSupply, enterprises can: ✅ Save millions in downtime and shipping delays ✅ Prevent stockouts, missed deadlines, and lost revenue ✅ Optimize decisions during volatile geopolitical or climate events ✅ Protect their brand from reactive crisis handling Here’s what we built: 🛰️ InfoGathering Agent Pulls real-time news, weather, geopolitical, and economic signals for every supply node using Tavily + LLMs. Always informed. Always scanning. 📡 Forecast Agent Predicts future disruptions using real-time data + memory context (Mem0). Built for high-precision, explainable forecasting. 🧪 Scenario Agent Simulates “what if” events (e.g., cyberattacks, port strikes, floods) using Monte Carlo modeling, cascade logic, and realistic probabilities. 📉 Impact Engine Visualizes chain reaction effects across supply networks. Understand which nodes fail, how fast, and what recovery looks like. 🛡️ Strategy Agent Suggests cost-effective mitigation strategies with ROI: rerouting, inventory buffers, alternate sourcing — with full implementation guidance. 🤖 Copilot Agent (UX) Built with CopilotKit — delivers an AI-powered planning interface that feels like working with a smart assistant. 📊 Plus: * Digital Twin Builder (React Flow) * Fully Schema-Validated Agent Outputs (Zod) * Memory + Retrieval via Mem0 * Supabase-powered backend * Clean, fast UI with Next.js + Tailwind 🧑💻 Built with ❤️ by Arnab Mondal Suman Jana Anirban Majumder and myself. ⚙️ Powered by: * Vercel AI SDK + Gemini 2.5 Pro * Tavily (search + crawl) * CopilotKit🪁 (AG-UI) (agent UX) * Mem0 (context memory) * Supabase (auth + data) * LangChain-inspired orchestration 🙌 Thanks to the Ariel Weinberger for an epic challenge — and for pushing the boundaries of autonomous AI. 📽️ Demo Video : https://lnkd.in/dD2FGH5u #100Agents #Hackathon #AI #SupplyChain #LangChain #CopilotKit #Tavily #Mem0 #DigitalTwin #MultiAgentSystem #AutonomousAgents #BusinessResilience #AIAgents #IntelliSupply #BuiltWithSupabase #Nextjs #OpenSourceAI