4497 - 使用 AllyCat 构建基于网络爬虫数据的 RAG 应用程序.pdf

1、IBM TechXchange2025 conference MCP and Agents:Open-Source Foundations for Agent Trust,Safety,and Scale1#IBMTechXchangeFred AraujuFred Arauju:IBM ResearchDean WamplerDean Wampler,Jeffrey BorekJeffrey Borek,Dave NielsenDave Nielsen:IBM and The AI AllianceSessions:2936,4497IBM TechXchange|2025 IBM Corp

2、orationOutlineContextForgeWhat Is the AI Alliance and Why?Building AI-empowered Applications:Agents,MCP,and MoreA Quick Look at Other AI Trends and AIA ProjectsFinal ThoughtsOutlineContextForgeWhat Is the AI Alliance and Why?Building AI-empowered Applications:Agents,MCP,and MoreA Quick Look at Other

3、 AI Trends and AIA ProjectsFinal ThoughtsOutlineContextForge MCP GatewaySecure Open-Source Gateway,Proxy and Registry for AI Agents and ToolsUnifying discovery,authentication,authorization,role-based access control,rate-limiting,observability,virtual servers,and protocol translation with a security

4、focus and configurable Agents use tools to accomplish complex tasksAI AgentA modular,goal-driven component that uses tools,memory,and reasoning to autonomously complete tasks via language modelsThese agents can plan,interact with APIs or other agents,and make decisions based on prompts,documents,or

5、user input.Theyre often orchestrated in a multi-agent system to collaboratively solve complex problemsThey typically use an agentic framework or orchestrator(e.g.,langgraph,langchain,autogen,crew.ai)for tool executionWhat can they do with tools?Retrieve external information(e.g.,web search,databases

6、,RAG)as necessary,call other Agents or request user inputGenerate artifacts(images,diagrams,word documents,excel files),interact with external APIs and systems using various Tools,run unit tests,etc.https:/ and tools turn“thoughts”into actionsLarge Language Models call Tools through Agentic Framewor