American forces are moving to close a persistent battlefield gap: enabling real time communication across multiple languages in operational environments.
The US Army and Navy are collaborating to develop a portable, AI-based translation tool to support speech-to-speech communication where connectivity is limited or unavailable.
The effort brings together the US Army Combat Capabilities Development Command Army Research Laboratory and the Office of Naval Research Global TechSolutions under a one-year agreement to develop a prototype known as ULTRA-MAR.
ULTRA-MAR enables real-time speech translation without network access and can operate in low-bandwidth or contested environments while handling domain-specific military terminology.
Testing is planned across Department of Defense units to validate performance in real-world conditions.

AI for Translation Support
The initiative reflects a broader push to move AI capabilities out of the lab and into operational use more quickly, particularly in areas where commercial tools remain limited or unsuitable for defense environments.
Built for secure, offline use, ULTRA MAR is developed in-house, allowing rapid modification to support new use cases, hardware platforms, and additional languages.
Once operational, the tool can enable direct communication between deployed personnel, partner forces, and local populations without relying on human interpreters.
The system also supports scenarios such as maritime communications, where sailors may need to interact with non-English speaking counterparts during exercises or patrols.
“We can deliver this capability quickly, because the specifications of our Army translator system already meet the strict requirements mandated by the Navy and Marine Corps,” said Dr. Stephen LaRocca, US Army computer scientist and principal investigator.
“Because the technology was designed specifically for use by Soldiers in contested environments, it already possesses many of the same defense-relevant features that Sailors and Marines need for their mission.”