Multilingual Video Translator
Upload a video and get a translated, dubbed version back. For this team hackathon project, I built the Express backend that connects transcription, translation, speech generation, and FFmpeg video processing.
- Express
- FFmpeg
- AssemblyAI
- ElevenLabs
Getting the stages to work together
I handled uploads with Multer, extracted audio with FFmpeg, and connected AssemblyAI transcription, translation, and ElevenLabs speech generation. The backend then remuxes the generated voice track with the original video. Server-sent events report progress while the browser waits for each stage.
Implementation references
What we shipped at LancerHacks
Our team won 1st Place in Access to Education at LancerHacks VII. My contribution was the backend and asynchronous media pipeline; the rest of the app was a team effort.
The prototype assumes English input and a fixed voice model. It also shares working files across jobs, so it needs changes before handling concurrent uploads reliably.