Multilingual Voice & OCR Translator

I want to bring everyday Hindi-English-Marathi conversations into one streamlined Android app. The idea is simple: I point the camera at a street sign or menu, the app grabs the text with Optical Character Recognition, instantly translates it, and then reads the result back to me through a clear Text-to-Speech engine. The same flow should work when I speak or type a phrase—I receive a fast, accurate translation plus an optional audio playback so I can mimic the correct pronunciation on the spot. Even for communication with auto rickshaw driver, sabzi mandi, hawkers etc. I want to make it as a communication tool Core flow • Capture text with OCR from photos or the live camera view • Translate bi-directionally between Hindi, English and Marathi • Convert translated text to natural-sounding speech A clean Material-based interface, offline fallback where practical, and smooth hand-offs between camera, translation, and audio are important to me; no clunky modal switches. Google ML Kit, TensorFlow Lite or any equally efficient on-device library is fine—use whatever delivers fast, lightweight performance without ballooning the APK size. Please supply a signed release-ready APK along with full Android Studio source so I can iterate later. A concise README that shows how to rebuild and replace API keys will round out the delivery.

Project ID: 40424929