LangGraph-Chatchat functions as a memory management plugin for the ChatChat conversational framework, utilizing LangGraph’s graph database model to store and retrieve conversation context. During runtime, user inputs and agent responses are converted into semantic nodes with relationships, forming a comprehensive knowledge graph. This structure allows efficient querying of past interactions based on similarity metrics, keywords, or custom filters. The plugin supports configuration of memory persistence, node merging, and TTL policies, ensuring relevant context retention without bloat. With built-in serializers and adapters, LangGraph-Chatchat seamlessly integrates into ChatChat deployments, providing developers a robust solution for building AI agents capable of maintaining long-term memory, improving response relevance, and handling complex dialog flows.

DeepSeek News Agent uses NewsAPI to fetch the latest articles matching user-specified keywords or topics and then processes them with OpenAI GPT models. It generates concise summaries, extracts sentiment scores, highlights trending themes, and clusters related stories. Users interact via a chat interface or command line to ask follow-up questions, request deeper analysis, or adjust filters. The agent maintains conversational context, allowing multi-turn queries for more refined insights, and can be extended or embedded into custom applications.