Unitra AI is a desktop application that translates voice chat, text, and on-screen game content in real time, displaying translations in a click-through in-game overlay. It combines on-device speech recognition with cloud-based translation optimized for gaming terminology, achieving roughly one-second end-to-end latency. The tool also supports streaming subtitles and is free to start with about two hours of daily voice translation.
Funding
Funding not disclosed
Founders
Product
Problem
Language barriers in online gaming prevent players from coordinating with teammates, streaming to international audiences, or engaging with players who speak different languages. Existing translation tools are not designed for real-time gameplay, requiring alt-tabbing, copy-pasting, or interrupting the action, which disrupts immersion and slows communication.
Solution
Unitra AI is a desktop app that translates game-related audio and text in real time, capturing both microphone and desktop audio simultaneously. It runs speech recognition locally on the device and sends the text to a cloud translation pipeline optimized for gaming vocabulary, delivering translated subtitles in about one second. Translations appear in a non-intrusive, click-through overlay that auto-detects the active game, so players never have to leave the action. The app also supports text and image translation for chat messages and on-screen UI text, and it can generate automatic subtitles for live streams. Audio privacy is a core feature: all audio processing stays on-device by default, with no uploads or recordings unless the user explicitly opts in.
Target Audience
Primary users are gamers who play with international teammates or stream to multilingual audiences, along with remote teams needing cross-language meeting support and language learners seeking immersive real-time translation.
Features
- Real-time voice translation with ~1 second end-to-end latency, combining on-device ASR with a cloud translation pipeline
- Dual audio source capture for simultaneous microphone and desktop audio, covering game voice chat, Discord calls, and meeting audio
- Click-through, always-on-top overlay with four display modes, auto-detecting running games and pinning translations on top
- Gaming-specific terminology engine that recognizes and correctly translates characters, items, abilities, and callouts unique to each title
- Privacy-first architecture: audio stays on-device by default, with no uploads, collection, or storage unless the user opts in
- Streaming support for automatic translated subtitles on live streams with no extra setup
- Text and image translation for pasted chat messages and captured on-screen UI text