Microsoft Launches Custom Neural Voice for Lifelike App Narration
Microsoft makes its Custom Neural Voice feature generally available through Azure, enabling developers to build highly realistic, humanlike text-to-speech for their applications. Strict access limits are in place to prevent the creation of misleading deepfakes.
Microsoft officially launches Custom Neural Voice, a powerful text-to-speech capability within Azure Cognitive Services that allows developers to create highly realistic, humanlike voices for their applications. By simply uploading recorded audio as training data, users generate a unique, natural-sounding voice tailored specifically to their brand or conversational interface.
Because the generated audio is so lifelike, Microsoft strictly limits access to this technology to ensure responsible use. This careful rollout is part of the company's broader commitment to ethical AI, aiming to protect individual rights, promote transparent interactions, and actively prevent the spread of harmful deepfakes or misleading content.
Major companies like AT&T, Progressive, and Duolingo already utilize this feature to enhance customer experiences. AT&T brings Bugs Bunny to life in its Dallas experience store, Progressive uses a custom Flo chatbot for customer inquiries, and Duolingo creates diverse character voices to make language learning more engaging for users.