Microsoft Translator Pro can translate foreign-language text found in images

Microsoft Translator Pro, a new app from the Redmond-based tech giant, is entering the public preview. With it, businesses can break language barriers with a range of features designed for the workplace.

Microsoft Translator Pro is a standalone, native mobile experience. People working in the same organization can use the app to communicate with one another or clients in different languages. In a video, Microsoft shows how a hotel receptionist speaks in English, a guest responds in Chinese, and the app takes care of the rest, translating the messages and playing them aloud.

The enterprise version of Microsoft Translator Pro includes a speech-to-speech translation feature. The app can simultaneously translate conversations and display the translations and transcriptions on the device. It can also work offline, allowing IT administrators to manage the app’s deployment and usage, view logs and conversation history, and configure settings to disable or export history to the cloud.

In a blog post, Microsoft says that the enterprise version of Microsoft Translator Pro offers a “high level” of translation quality and security. Once an organization’s IT admin grants access, employees can sign in to the app using their organization’s credentials. Microsoft says conversational data stays within an organization’s Azure tenant and isn’t accessible to Microsoft or other parties.

In addition to the new version of Microsoft Translator Pro, Microsoft is rolling out a feature in its Document Translation service in Azure AI Translator that processes text embedded within images within scanned documents.

The Document Translation service in Azure AI Translator now leverages the Microsoft Azure AI Vision API to extract, detect, and translate text from images. Microsoft says the service works with various file types, including PDFs and photos, and it’s available in preview.

Customers must opt into the feature when creating a translation request to translate text in images. A new parameter called translateTextWithinImage is now offered under “options” in the Azure AI Translator API. When the feature is enabled, the response will include details about the number of image scans that succeeded and failed.

Microsoft says that the feature, which arrives in January 2025, will be available in the Azure AI Services resource and incur charges based on usage.

Speaking of Azure, the Redmond-based tech giant has unveiled many new capabilities, platforms, and enhancements coming to the service, including Azure Local, Azure AI Foundry, and Azure AI Content Understanding.