A few weeks back, I shared our vision for the agent era: agents should both run on HubSpot and run HubSpot itself. Now I want to dig a bit deeper into what it really means to “run HubSpot,” and walk through our latest move toward making that vision real. Today, it’s not just employees logging into HubSpot to get work done — it’s agents. And those agents need to operate as effectively as possible on your behalf, no matter where they’re running. That last point is critical. An agent doesn’t live in a single interface or a single infrastructure. With AI Connectors, HubSpot context and actions are already accessible from Claude, ChatGPT, and other tools where teams spend their time. Now, we’re extending that agent infrastructure to a new surface: the Command Line Interface (CLI). Introducing the HubSpot Agent CLI The HubSpot Agent CLI brings HubSpot’s data and intelligence directly into the environments where GTM and ops teams design their own workflows — Codex, Claude Cowork, and Claude Code — enabling agents to automate repetitive, large-scale, and scheduled tasks. The easiest way to think about it: take the questions you keep asking or the tasks you repeatedly perform in chat, and turn them into automations. Create automations in Codex or schedule them in Cowork, and the work gets done automatically before you even sit down at your desk. It’s built on the same core layer as our public APIs and the MCP server that already power our AI Connectors — and it’s meant to work alongside them, not replace them. AI Connectors are…