HighByte 为 AVEVA PI System 提供 MCP:智能体拿受治理的工业数据而非裸标签
HighByte 介绍 Intelligence Hub 中的 MCP Server:可将 PI Data Archive 与 Asset Framework 的点位、资产、事件框等经可配置管道暴露为 MCP 工具,合并 MES 与 ERP 等上下文,并用令牌控制代理访问。
工厂数据总线开始给智能体发受治理的工具,而不是整库标签。
HighByte 说明:其 Intelligence Hub 里的 MCP Server,可以把 AVEVA PI System(含 PI Data Archive 与 PI Asset Framework)的点位、资产、事件框、数字状态表等,经无代码管道与 MES、ERP、LIMS 等上下文合并后,暴露成带名称、描述与参数的 MCP 工具,供 AI 智能体调用。访问用基于令牌的鉴权,并可监控代理与工具的交互,避免直接暴露原始标签或引发权限冲突。
对制造智能化,这意味着预测维护、批次追溯、值班问答类代理可以读到带上下文的 PI 数据,同时把权限留在工业数据平台侧。实施要点:先按资产树最小化暴露工具面,再为只读分析与有限写回分令牌,并审计工具调用。把 MCP 工具目录纳入变更管理,避免产线人员私自扩大点位范围。
对行业读者而言,关键是核对官方文档中的适用范围、权限模型与发布时间表,避免把预览能力误当成全量通用能力。
对行业读者而言,关键是核对官方文档中的适用范围、权限模型与发布时间表,避免把预览能力误当成全量通用能力。
HighByte MCP Server Exposes Governed AVEVA PI Data to AI Agents
Q: How do you give AI agents access to data from @AVEVA PI System without exposing raw tags or creating security conflicts?
A: The #MCP Server available in #HighByte Intelligence Hub allows PI System datasets to pass through configurable MCP tools for agents to call. Use the Intelligence Hub to:
✅ Read PI points, assets, event frames, digital state tables, and more from PI Data Archive and PI Asset Framework ✅ Merge PI data with context from MES, ERP, LIMS, and other systems in code-free pipelines ✅ Expose custom pipelines in the Intelligence Hub as MCP tools, each with a name, description, and parameters that tells agents when and how to use them ✅ Control which agents can use which tools with token-based authentication, and monitor how they interact
The result: AI agents get contextualized, governed PI System data through the MCP Server, and your team stays in control of what they can access.
More here 👉 https://t.co/GMVuSNpx45