Skip to main content
When you communicate offline with foreign colleagues, customers, or partners, AI Minutes can translate both sides of the conversation into the languages each person needs in real time, presenting the result as subtitles or audio. This helps remove cross-language communication barriers and makes face-to-face conversations more efficient and natural.

​
Scenario Highlights

​
1.1 Bilingual Split Screen: Each Person Views One Half

After face-to-face translation is enabled, the phone automatically splits into upper and lower areas. You see the translated text in the other person’s language, such as Chinese, while the other person sees the translation in your language, such as English. Both people can view the translation at the same time. This is suitable for close-range communication at meeting tables, cafes, and similar scenarios. When switched to landscape mode, the screen is automatically split into left and right areas. It can also be projected to a meeting room display, making it suitable for business discussions and other meeting scenarios.

​
1.2 Voice Broadcast: Choose as Needed

Voice broadcast of translation results is supported for different usage habits:
  • Without headphones: when you speak Chinese, the system plays English through the speaker for the other person. When the other person speaks English, the system plays Chinese for you.
  • One person wearing headphones: you listen to the translation of the other person’s language, such as Chinese, through headphones, and the other person will not hear the voice broadcast.
  • Two people wearing headphones (two-person simultaneous interpretation mode): you wear the left earbud to hear Chinese, and the other person wears the right earbud to hear English, creating a more private interpretation experience.

​
1.3 Automatic Language Recognition and Multiple Language Support

Chinese, English, and Japanese automatic recognition and translation are currently supported. The system can intelligently determine which language the speaker is using and output the target-language translation in real time, without manual language switching.

​
1.4 Conversation Records and Minutes on Demand

After the conversation ends, the system generates AI Minutes minutes with a complete bilingual transcript, making it easier to review key information later. Note: the recording state is shown less prominently in the interface to avoid affecting customer willingness to communicate.

​
Operation Flow

​
1. Entry Point for Face-to-Face Translation

​
From the AI Minutes Home Page

On the AI Minutes home page, open the plus entry and click the Face-to-face translation card to enter the translation interface. image.pngimage.png

​
2. Set Target Languages

Select the translation languages in the upper-right corner of the interface:
  • The language you want to see, such as Chinese.
  • The language the other person wants to see, such as English.
image.png

​
Common Q&A

Q: Which languages are currently supported? A: Chinese, English, and Japanese mutual translation is supported, and the spoken language can be recognized automatically. Q: Will voice broadcast feel intrusive? A: Private listening with headphones is supported. When using speakers, the system uses half-duplex logic: it plays the translation after one person finishes speaking, and the other person speaks after listening, helping avoid overlapping audio. Users can also turn off voice broadcast and view subtitles only. Q: Can this be used for conferences or multi-person meetings? A: Landscape projection to a large screen is supported, allowing multiple people to watch bilingual subtitles. For large meetings, DingTalk Meetings + real-time subtitles is recommended for better results. Q: How accurate is the translation? A: For mainstream Chinese, English, and Japanese scenarios, translation is powered by the Alibaba Cloud large-model translation engine and is generally accurate. Recognition accuracy depends on original audio clarity; dialects and strong background noise may affect results.