Claude users can now call up interactive apps directly within the chatbot interface. This new feature, announced by Anthropic on Monday, brings a selection of workplace tools into Claude. The initial launch apps include Slack, Canva, Figma, Box, and Clay, with a Salesforce implementation expected soon.
For each enabled app, a logged-in instance of the service becomes accessible to Claude. This allows users to perform tasks like sending Slack messages, generating design charts, or accessing cloud files without leaving the conversation.
Anthropic explained the reasoning behind the feature in a blog post, stating that analyzing data, designing content, and managing projects all benefit from a dedicated visual interface. The company believes that combining these interfaces with Claude’s intelligence will let users work and iterate faster than either could offer alone.
Access to this new feature is limited to Pro, Max, Team, and Enterprise subscribers and is not available to free users. Eligible subscribers can activate the tools through the Claude directory.
This system is similar to the Apps feature launched by OpenAI in October. Both integration systems are built on the Model Context Protocol, an open standard introduced by Anthropic in 2024. The MCP launched its own support for apps in November, incorporating work from both companies.
The new apps are expected to become particularly powerful when integrated with Claude Cowork, an all-purpose agent tool launched by Anthropic last week. Built on Claude Code, Cowork allows users to assign complex, multi-stage tasks that use large datasets. When combined with the apps feature, Cowork could be granted access to cloud files or ongoing projects to perform actions like updating a marketing graphic in Figma or using new data from a company’s Box instance. While apps are not available in Cowork at launch, Anthropic confirmed this integration is coming soon.
Anthropic acknowledges that agentic systems can be unpredictable. The company’s own safety documentation for Cowork encourages users to monitor the agent closely and avoid granting unnecessary permissions. It specifically recommends caution when granting access to sensitive information like financial documents or personal records, and suggests creating a dedicated working folder for Claude rather than providing broad access.