Tensor AI
Open Chat

ChatVLM Now Speaks MCP

ChatVLM now supports the Model Context Protocol (MCP) — the open standard for connecting AI assistants to external tools and data sources.

What this unlocks

MCP servers give ChatVLM new capabilities the same way extensions give a browser new powers. Connect a server, and every conversation can use it:

  • Your data — internal knowledge bases, databases, ticketing systems
  • Your tools — anything that exposes an MCP interface, from developer utilities to SaaS products
  • Your workflows — chain ChatVLM's built-in tools (search, code interpreter, documents) with your own

Because MCP is an open standard, the ecosystem is already large and growing daily. If a tool speaks MCP, ChatVLM can use it — no custom integration work on our side or yours.

Working with the Smart Router

Connected tools compose naturally with what makes ChatVLM different: the Smart Router still picks the right model for every turn, and now it hands that model your tools as well. A data-heavy question can route to a model that excels at tool use; a creative one routes elsewhere — your connections work everywhere.

Getting started

Head to settings, add an MCP server, and start asking questions that touch your own systems. Connections are scoped to your account, and — as with everything on ChatVLM — requests run under our contractual zero-data-retention guarantee.

This is the first step in a broader push to make ChatVLM the place where your AI and your systems meet. More on that soon.