Inspiration
I wanted to make watching foreign-language videos easier, especially for language learners who want to understand both the original speech and its translation.
What it does
LinguaEcho is an AI-powered video and audio translator. Users can upload media, translate speech into different languages, listen to bilingual playback, follow synchronized transcripts, and export translations.
How we built it
I built the frontend using React, TypeScript, and Tailwind CSS, with Python Flask for the backend. I integrated OpenAI APIs for transcription, translation, and text-to-speech. The app is hosted using GitHub Pages and Google Cloud Run.
Challenges we ran into
The biggest challenge was synchronizing the original video with AI-generated speech, ensuring the video pauses during translations, and handling accurate sentence timing.
Accomplishments that we're proud of
I'm proud of building a working multilingual translator with four playback modes, interactive transcripts, and a live website.
What we learned
I learned how to integrate AI APIs, apply prompt engineering techniques like Persona Pattern and Few-Shot Prompting, synchronize audio playback, and deploy a full-stack application.
What's next for LinguaEcho
I plan to improve translation accuracy, support more advanced audio synchronization, and add browser-local storage so users can keep their projects.
Built With
- ai
- audiotranslator
- chatgpt
- flask
- fullstack
- githubpages
- google-cloud
- languagelearning
- multilingual
- openai
- promptengineering
- python
- react
- speechtotext
- texttospeech
- translation
- translator
- typescript
- videotranslator
- webdevelopment
- whisperai
