I’m building a static frontend hosted on GitHub Pages for a video dubbing tool that requires speech-to-text, translation, TTS voice cloning, and lip-sync ML models. Since GitHub Pages doesn’t support any backend, I need to host all ML models separately and call their inference APIs from the static UI. ✅ What I’ve Done Created a React/HTML/CSS frontend and deployed it on GitHub Pages. Broke my model pipeline into multiple REST endpoints: ASR, translation, voice cloning, lip-sync. Tested them locally with Flask/Gradio/FastAPI. Core Issue I need a free (if possible, with GPU service for fast processing) hosting solution for the ML APIs that I can reliably call from my GitHub Pages site. It must support Python ML frameworks (e.g., Whisper, MarianMT, TorToise-TTS, Wav2Lip) and be accessible via HTTPS. Continue reading...
