Key Highlights
- Anthropic has launched interactive apps inside Claude, starting with workplace tools
- Supported apps include Slack, Canva, Figma, Box, and Clay
- The feature is available only to paid Claude users, not free accounts
- The system is built on the open Model Context Protocol (MCP)
Anthropic has launched interactive apps inside its Claude chatbot, allowing users to access workplace tools like Slack, Figma, and Box directly within the AI interface. The update matters because it turns Claude from a text-based assistant into a hands-on work companion that can interact with real tools, data, and files in real time.
The new feature is rolling out to Pro, Max, Team, and Enterprise subscribers and is currently unavailable to free users. Eligible users can enable apps through Claude’s directory, marking a major step in Anthropic’s enterprise-focused AI strategy.
What Are Anthropic’s New Claude Apps?
With this update, Claude users can open interactive apps inside the chatbot itself. Once logged in, Claude can access enabled services and perform tasks based on user instructions. For example, users can ask Claude to send Slack messages, pull files from Box, generate visuals in Canva, or work on designs in Figma.
According to Anthropic, the goal is to combine Claude’s reasoning abilities with visual and task-specific interfaces. This approach allows users to analyze data, design content, or manage projects without switching between multiple platforms.
According to Anthropic, the goal is to combine Claude’s reasoning abilities with visual and task-specific interfaces. This approach allows users to analyze data, design content, or manage projects without switching between multiple platforms.
Which Workplace Tools Are Supported at Launch?
At launch, Anthropic is supporting Slack, Canva, Figma, Box, and Clay. A Salesforce integration is also in development and expected soon. Each app runs as a secure, logged-in instance that Claude can access only after user approval.
This design ensures Claude can act on real workplace environments while still requiring explicit permissions. As a result, users maintain control over what the AI can see and do.
How Is This Different From Other AI App Integrations?
The system closely mirrors OpenAI’s Apps feature, introduced in late 2025. Both platforms rely on the Model Context Protocol, or MCP, an open standard introduced by Anthropic in 2024 to help AI models interact safely with external tools.
By using MCP, Anthropic allows third-party apps to plug into Claude without custom integrations. This makes it easier for more enterprise tools to join the ecosystem over time.
What Role Does Claude Cowork Play?
The new apps are designed to work alongside Claude Cowork, Anthropic’s recently launched agentic AI tool. Cowork allows users to assign multi-step tasks across large datasets. While app support is not live inside Cowork yet, Anthropic has confirmed that deeper integration is coming soon.
However, Anthropic has also warned users to be cautious. Its safety documentation advises against granting access to sensitive data and recommends using dedicated folders instead of broad system permissions.
In the conclusion, Anthropic’s latest update shows how interactive apps could reshape how agentic AI fits into daily work, while reinforcing the need for careful oversight as these systems gain more autonomy.
