Ali AI coding assistant Qoder officially launched the real-time voice interaction agent Qoder Voice, driven by the full-duplex voice model Qwen-Audio-3.0-Realtime. It supports developers to complete task creation, solution discussions, and code execution through natural voice, achieving real-time collaboration between voice and AI coding.

According to the introduction, users can click on the voice entry in the main interface of Qoder Quest mode, and then call Qoder Voice through a floating window at any interface. After issuing a voice command, the system will automatically create a task, call relevant tools, and execute them in the background. Users do not need to wait for the results, and can interrupt, ask follow-up questions, or adjust requirements at any time during the execution process, achieving synchronized voice interaction and task execution.

QQ20260728-172605.jpg

For complex development tasks, Qoder Voice supports multi-turn discussion-based interaction. The intelligent agent will first generate an execution plan, and users can supplement requirements or change directions through voice. The system will then optimize and continue executing based on feedback in real time. For example, in a code refactoring scenario, users can request a specific plan and add features such as caching, and Qoder Voice will immediately adjust the execution strategy.

Currently, Qoder Voice has been launched on the international version of Qoder, and is available for individual subscriptions, personal trial versions, and enterprise subscriptions, with a 3-day free trial offered. Alibaba stated that the Chinese version of Qoder (Qoder CN) will also be launched soon.

As large models, multimodal technologies, and AI programming tools continue to integrate, voice interaction is becoming a new access point for developers to collaborate with intelligent agents. The launch of Qoder Voice further promotes the evolution of AI programming from traditional text input to more natural and efficient real-time voice interaction.