Global artificial intelligence (AI) product and research corporations DeepL said on the 3rd that it officially launched DeepL Voice API, a voice translation solution that integrates real-time speech recognition and translation into applications.
DeepL Voice API lets developers process audio via streaming to receive speech recognition results in the source language and implement real-time translation into up to five target languages. Through this, corporations are expected to lower language barriers in customer service and internal communications and build a smoother multilingual communication environment.
The API is expected to see wider adoption in industries where voice communication is core, such as contact centers and business process outsourcing (BPO). By adding real-time speech recognition and translation to existing systems, it enables immediate responses in multilingual customer service without transferring to another agent or switching to written communication.
On the operations side, DeepL Voice API provides clear speech recognition and translation to support customer service quality control and agent training. This shortens review times, enables fair evaluations across regions, and helps more systematically identify agent performance and knowledge gaps. It also reduces inefficiencies caused by language barriers—such as longer call duration, repeat inquiries, and extra expense from misunderstandings—contributing to an overall improvement in customer experience.
Gonçalo Gaiolas, DeepL chief product officer (CPO), said, "Language barriers in customer service can directly lead to negative experiences," and added, "By providing tools that enable real-time communication in any language, customer support organizations can shift from cost-incurring departments to value-creating organizations through customer satisfaction." He went on to say, "Voice API will reduce the burden of agent transfers or workaround solutions and boost on-site operational efficiency."
DeepL Voice API supports staffing based on expertise, reduces reliance on language-specific personnel, and contributes to optimizing operating expense and expanding the talent pool. It also enables stable service delivery during periods with a shortage of specialized language staff, such as nights, weekends, and holidays. Agents can use real-time speech recognition results, caption translation, and audio interpreting together to communicate with customers more naturally and confidently.
The launch also includes an early access program for the Voice-to-Voice interpreting feature, which will run for six weeks starting in mid-February. This allows agents to communicate with customers in their preferred language in real time while providing simultaneous voice interpreting.
Starting today, DeepL Voice API is available to DeepL API Pro subscribers, and corporations can access the service by checking the API documentation or contacting the sales team.