On May 7th, Qwen launched AI voice input capability on PC, allowing users to directly use it in various desktop applications through shortcut keys. Qwen's voice input supports removing filler words, correcting errors, and formatting spoken content, and can intelligently respond based on the context. It can also directly issue commands for writing, answering questions, and translating.

Qwen's voice input demonstrates strong semantic parsing capabilities. For example, given a spoken passage with speech errors: "Boss Wang, the meeting is scheduled at two o'clock... no, three o'clock, the location is the old meeting room, remember to bring that... market research report." Qwen can filter out redundant words such as "um" or "that" in real time and correct speech errors, directly outputting structured text: "Boss Wang, the meeting is scheduled at 3 p.m., the location is the old meeting room, please bring the market research report."

4d105a28e0a11d7d174fd97a2192c2eb.jpg

Reconstructing colloquial expressions is not limited to text simplification, but also includes logical error correction and structured organization. When users are writing weekly reports or minutes, even if their expressions are logically跳跃, Qwen can perform semantic summarization based on the context. For example, if a user casually speaks about work progress, Qwen can automatically organize it into a structured weekly report with clear points.

When working or creating, users can use Qwen through voice commands without leaving the editor. For example, if you want to insert last year's GDP data in the text, just say "Insert the national GDP data for 2025 for me." When browsing websites or reading professional English papers, if there is something unclear, select the text and say "Explain it for me" or "Translate it for me," and Qwen will automatically explain or translate.

In scenarios like DingTalk, WeChat, or emails, users can also ask Qwen to generate replies based on the context. For example, when replying to an English email from a client, users can simply verbally state their requirements in Chinese, and Qwen can generate a correctly formatted English email reply in the input box.

Currently, the Qwen voice input function is fully open, and all users can freely use this feature through the Qwen PC client.