Claude Now Lets You Run Apps Directly Inside Chat

Anthropic has rolled out a new extension to its Model Context Protocol that lets Claude interact directly with third-party apps inside the chat itself. Not summaries. Not suggestions. Actual interaction, with support for tools like Slack, Asana, Figma, and a growing list of others.

Instead of jumping between tabs or copying instructions back and forth, users can now see app content appear inside Claude and take actions from there. Tasks can be reviewed. Messages can be referenced. Design files can be pulled into the conversation. The chat becomes the place where work happens, not just where it gets discussed.

What’s Actually New Here

• This builds on MCP, which Anthropic introduced as a way for Claude to securely connect to external tools and data sources.
• Earlier versions focused on letting developers wire Claude into systems behind the scenes.
• This update makes that connection visible and interactive for everyday users.

The key difference is execution. Claude is no longer just reading from connected apps or generating instructions about them. It can now act within those environments through authenticated connections, using the same permissions a user already has.

Anthropic is positioning this primarily for teams and work settings, where collaboration tools already define the flow of the day. The feature is rolling out gradually, starting with paid plans and enterprise use cases.

A Short Bit of Context

This idea isn’t brand new. Workflows have been inching toward consolidation for years. Slack bots, browser extensions, automation tools, and low-code platforms all tried to reduce how much context switching people do. Most of them helped, but none of them became the place where everything met.

What’s different now is that chat has become a comfortable default. People already think out loud there. They already paste drafts, links, and half-formed ideas into a message box. Letting apps live in that same space feels less like a leap and more like a next step.

What This Changes

The practical benefit is obvious. Fewer tabs. Less back-and-forth. A smoother way to move from thinking to doing. For a lot of people, that alone will be enough.

The more interesting part is what it turns chat into. Claude is no longer just a helper sitting beside your tools. It becomes a coordination layer that touches many of them at once. That raises questions about permissions, oversight, and how comfortable users are with an AI acting inside their work systems rather than commenting from the outside.

Nothing about this feels huge. And that might be the point. Big changes in how work happens rarely announce themselves. They usually arrive as small conveniences that slowly become hard to imagine working without.