DeepL Voice

What it does

DeepL Voice takes in live speech from a microphone or a meeting audio stream and returns real-time translation as voice, live subtitles, or both. Speakers can choose from more than 40 languages, and the output keeps each speaker's voice character where voice-to-voice translation is used.

It has three forms. Voice for Meetings adds translated audio and subtitles inside Microsoft Teams, Zoom Meetings and Google Meet. Voice for Conversations is for face-to-face talks, such as frontline staff serving customers, using on-device translation on iOS, Android and the web. The Voice API embeds speech translation in contact-center and other business tools. Its distinguishing controls are glossaries, Spoken Terms for company-specific words and acronyms, and admin-shared style profiles.

It suits international teams, customer-facing staff and developers building multilingual voice channels. A web demo needs only a microphone, while Meetings requires a DeepL Voice for Meetings subscription. Voice pricing details were not visible, and the full language list is held in a separate help page. Meeting data is processed temporarily in memory and not used to train models.

What it can do

• Voice-to-voice translation in online meetings
• Live translated subtitles in Microsoft Teams, Zoom Meetings and Google Meet
• On-device in-person conversation translation, one-on-one or in groups
• Glossaries, Spoken Terms and shareable style profiles for terminology
• Voice API for speech-to-text and speech-to-speech with low-latency streaming
• SSO (OIDC and SAML), MFA, role-based permissions and audit logs

Works with

• Microsoft Teams
• Zoom Meetings
• Google Meet

Example workflows

Multilingual team meeting

Input: a scheduled Teams, Zoom or Google Meet call with participants speaking different languages. Steps: admin sets up Voice for Meetings, shares a style profile with glossaries, participants pick their language. Output: translated audio and live subtitles for each participant.

Frontline customer conversation

Input: a staff member and customer speaking different languages. Steps: open the DeepL app or web page, select both languages, start speaking. Output: on-device translation shown and spoken for each side of the conversation.

Before you choose

• Voice for Meetings requires a separate subscription; Voice pricing was not confirmed.
• Admins can enforce shared terminology with style profiles, which matters for technical or regulated settings.
• Meeting data is held temporarily in memory and conversation data stays on-device, with no model training on customer data.

Listed under

• Productivity
• Meetings and notes
• Contact centers
• Customer service
• Enterprise
• In-person
• In-person conversations
• Interpretation

Get started

Meetings requires a Voice for Meetings subscription; enterprise via sales.

1. Try the web demo by selecting languages and choosing Start speaking with your microphone.
2. For meetings, subscribe to DeepL Voice for Meetings and enable it in Teams, Zoom or Google Meet.
3. For API use, get a DeepL API key and follow the Voice API quickstart in the developer docs.

The Individual plan includes 30 min of DeepL Voice per user/month, but dedicated Voice plan prices were not confirmed.

We couldn’t find published prices for DeepL Voice. See their pricing page ↗

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• DeepL Voice: instant, secure voice translation for global teams ↗www.deepl.com
• DeepL Pro | Translate Text, Word Docs & Other Docs Securely ↗www.deepl.com
• DeepL for Developers ↗www.deepl.com
• Empower your enterprise workflows with Language AI | DeepL ↗www.deepl.com
• What MCP changes about enterprise software ↗www.deepl.com

Company website

deepl.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.