Cursor and Google introduce two-way voice interaction for AI coding agents

New Voice Features for AI Coding Agents

Cursor and Google have both released new features that allow developers to use two-way voice communication with their AI coding agents. Cursor's "Voice Chat with Agents" uses Grok Voice, while Google's "Voice mode" in Antigravity integrates Gemini Live. These updates enable users to speak instructions and questions to their agents and receive verbal responses and updates without typing.

Addressing Developer Workflow

The introduction of voice interaction aims to address the current rhythm of working with coding agents, which typically involves typing prompts, waiting for the agent to process, reading the output, and then typing again. Google's developer relations engineer, Nick Van der Auwermeulen, noted that much of engineering involves verbal communication, such as discussing bugs with teammates or thinking out loud. The new voice modes seek to replicate this dynamic with AI agents.

Continuous Interaction

Both Cursor and Google highlight the benefit of continuous voice interaction. Google's Antigravity voice mode maintains an open call while the agent works, allowing developers to hear updates or approve commands without needing to return to the keyboard. Similarly, Cursor's feature allows users to describe changes, ask about ongoing tasks, and receive updates verbally as the work progresses.

Antigravity Platform Overview

Antigravity is Google's agentic software development platform, which debuted with Gemini 3. It received an update to version 2.0 in May, including a separate desktop application. The platform also offers a terminal client, Antigravity CLI, and extensions for integration with development environments like VS Code, JetBrains IDEs, and Xcode.