CometChat Launches Calling Skills for AI Agents, Expanding Integration Tools

CometChat has announced the launch of Calling Skills for AI Agents, the second iteration of its AI agent integration platform. The new offering enables developers to add HD voice and video calling capabilities to applications across six major platform families—React, Next.js, React Native, Angular, Android, iOS, and Flutter—with minimal setup time.

The Calling Skills feature, included in version 4.2.0 of the CometChat Skills package, allows developers to integrate calling functionality using a single command (npx @cometchat/skills add) or through UI Kits and SDKs. The product includes advanced features such as call recording, screen sharing, call logs, raise hand functionality, broadcast mode, and picture-in-picture capabilities. CometChat designed the tool to scale with high bandwidth demands while maintaining compliance with global standards.

A key distinguishing feature of Calling Skills is its intelligent setup process, which asks developers a critical upfront question: "Ringing or Session?" This choice determines the integration path. Ringing mode provides a full incoming/outgoing call surface with native platform support like CallKit on iOS and ConnectionService on Android, while Session mode enables link-driven meeting rooms where participants join the same session ID without requiring a ringing surface. After this determination, the system automatically detects the developer's framework and SDK version, scaffolds the correct file structure, and runs a 23-point verification pass to catch configuration issues across platforms.

The launch represents an extension of CometChat's earlier Chat Skills product, which launched a few months prior and allows AI agents to integrate chat functionality in approximately 15 minutes. The Calling Skills integration is fully compatible with existing Chat Skills implementations and requires no changes to previous chat integrations. According to CometChat's engineering team, the tool works seamlessly with popular AI development platforms including Cursor, Kiro, VS Code Copilot, Claude, and Codex.

The announcement has drawn attention to the broader challenge of SDK maintenance across multiple platforms. Community feedback highlighted concerns about how the tool handles API drift and breaking changes that occur independently across the six supported platform SDKs over time, suggesting this remains an ongoing consideration for developers evaluating long-term integration stability.