About DeepL Voice for Meetings

Voice for Meetings is an effortless way to bring accurate, real-time translation into your virtual meetings without requiring a dedicated app or a complex setup process.

Key benefits & value

How it works

1. Connect a meeting (DeepL Voice for Meetings license required) 
   Go to the DeepL website and enter the invitation URL of the meeting you want to translate. 
2. Admit the DeepL bot
   DeepL invites a bot to your meeting. Once the bot joins the meeting, it begins streaming audio for translation and sends a join link to all participants via the meeting Chat.
3. View captions
   Participants follow the link to open a dedicated DeepL browser window where they can select their preferred meeting language (to speak and to read captions). 

Requirements

• You need a DeepL Voice for Meetings license to connect a meeting
• Your organization must allow external participants to join your meetings
  This is required in order for the DeepL bot to join the meeting to stream audio for translation.
• The Chat function must be enabled
  Once the DeepL bot joins the meeting, it sends a link in the Chat for participants to access meeting translation
• (Optional) Add deepl.com to your trusted domains
  This prevents your organization from blocking the DeepL bot from joining meetings and ensures that the DeepL bot's details are correctly identified and displayed (e.g., name, profile picture).

Meeting types

Available languages

See the full list of available languages - DeepL Voice languages. 

Available features

• Pause translation
  The pause translation feature removes the bot from the meeting, so users can be reassured that no audio is being streamed when sensitive topics are being discussed. Disabling Pause translation will prompt the bot to rejoin the meeting.
• Formality
  Select the tone that best matches your meeting context. Choose between the following options:
  • Formal/polite (default)
    This is best for formal settings, e.g., a meeting with clients
  • Informal/plain
    This is best for casual settings, e.g., a meeting with your team
• Download transcript (not available with the Core plan)
  Download a transcript of your meeting. Learn more - Download meeting transcript. 
• Glossaries (not available with the Core plan)
  Apply a glossary to specify how certain words or phrases should be translated for all users. Learn about the glossary feature - glossary.
• Spoken terms (not available with the Core plan)
  Apply spoken terms to specify how certain words should be transcribed in a meeting. Learn more - About spoken terms. 

Data security

Access a detailed Voice data flow to support your security and data assessments in the Documents section of DeepL Trust Center.

DeepL guarantees the following:

• DeepL does not permanently store meeting transcription or translation data
• DeepL temporarily stores the latest messages in memory until the end of the call
• Data is deleted from memory once the call is terminated
• Transcriptions and translations accessed by users are stored on their local device and will persist for maximum 30 days (admin-controlled period) 
• Customer data is not used to train DeepL's language models

Sub-processors

Recall.ai

Unified experience is supported by a dedicated meeting-bot service provided by Recall.ai. Recall specializes exclusively in meeting-bot integrations, enabling DeepL to focus on delivering the core translation experience while leveraging Recall's industry-leading reliability and security. The business entity behind Recall.AI — Hyperdoc Inc. — is included as a sub-processor under the DeepL Voice Data Processing Agreements. The service operates from EU-based data centers and is contractually bound to a strict zero-data-retention policy, fully aligned with DeepL's longstanding high privacy standards. 

Learn more about Recall's robust security and compliance posture by visiting their Trust Center. 

Speechmatics

Speechmatics power the transcription of the languages listed in DeepL Voice languages. Speechmatics have enterprise-level security and a zero-retention agreement with DeepL, therefore, once processed, all transcription data is deleted from their servers once processed. Learn more about their security posture by visiting their Trust Center. 

Limitations

• DeepL Voice for Meetings supports up to 300 participants
• DeepL Voice does not support meeting translation for breakout rooms