タグ: Whisper speech recognition
2026/08/06Hey, are you Earthlings having trouble with Earth-speak?
OBS + Cloudflare Workers AI + AI Gateway Subtitle Script v6.3 is a free, open-source Python tool designed for real-time speech recognition and subtitle translation in OBS Studio. Powered by Cloudflare Workers AI and AI Gateway, it leverages OpenAI's Whisper model for speech-to-text and Meta's Llama 3.1 8B Instruct for instant translation (supporting Japanese, English, and auto-detection). By utilizing Cloudflare's free tier (10,000 Neurons/day), prompt caching, and low-latency API routing, this script allows live streamers and VTubers to display high-accuracy, translated subtitles without incurring monthly fees or GPU hardware overhead. Key features include dynamic noise gate filtering, frequency-based Voice Activity Detection (VAD), hallucination suppression, non-blocking audio pipelines, and seamless text file output for OBS Text (GDI+) sources.