Mixhalo is now integrated into the DeepL platform, bringing breakthrough ultra-low latency audio technology to voice translation and setting a new standard for what real-time multilingual communication can achieve.
Why Mixhalo?
One of the biggest technical challenges in real-time voice translation is latency. When a speaker’s words are translated and delivered to listeners, any perceptible delay disrupts the natural flow of conversation, making exchanges feel stilted and unnatural. Mixhalo possesses industry-leading ultra-low latency audio transmission technology, with delays measured in tens of milliseconds—well below the threshold of human perception.
To put this in context: the average human reaction time to auditory stimuli is approximately 150 milliseconds. Mixhalo’s technology operates at a fraction of that, meaning the translated audio reaches listeners before their brains register any gap. This isn’t just a technical achievement—it’s the threshold at which real-time voice translation stops feeling like translation at all and starts feeling like natural conversation.
Technical Integration: Building the Fastest Translation Pipeline
By combining Mixhalo’s audio transmission technology with DeepL’s language AI engine, we have created an end-to-end voice translation pipeline that is faster than any other solution on the market. From audio capture, language identification, and neural translation to speech synthesis and transmission—the entire process has been optimized for real-time communication.
The integration spans the entire stack. Mixhalo’s audio transport layer handles the capture-to-delivery path with sub-200ms total pipeline latency. DeepL’s translation engine processes the recognized speech with the same neural models that achieved a 94% win rate in independent blind tests against Google Translate, OpenAI GPT-5.2, and Anthropic Claude Opus-4.6. The result is not just fast translation, but fast translation at the highest quality benchmark in the industry.
What This Means for the Enterprise
For global enterprises, this means meetings can truly happen in real time. No more awkward pauses, no more missed speaking opportunities, no more translation subtitles that lag behind the speaker. Whether it’s a cross-border boardroom negotiation or a safety briefing on the factory floor, every participant experiences immediate, natural translation.
The impact on decision-making speed is profound. When executives and teams can communicate in real time without language friction, the latency of organizational decision-making itself decreases. Urgent issues that previously required sequential translation and clarification can now be resolved in a single conversation. The technology doesn’t just translate words faster—it accelerates the entire business cadence of global organizations.
Platform-Wide Impact and Massive Scale
Mixhalo’s technology also extends the scalability of DeepL Voice. We can now support meetings of unprecedented scale—thousands of simultaneous participants—while maintaining low-latency delivery for every individual listener. This is a breakthrough for large enterprise town halls, global all-hands meetings, and virtual events where language barriers have historically limited participation.
We are integrating this technology across Voice for Meetings, Voice for Conversations, and the Voice API, ensuring that you get the fastest voice translation experience no matter how you use DeepL Voice. The same ultra-low latency pipeline that powers a two-person conversation also scales to a 5,000-person company-wide broadcast, with every participant receiving translated audio in their preferred language at the same sub-200ms pace.