Stream Transcription by javi

Watch a Twitch stream, a YouTube video or a call in a language you don't speak, and read it in yours, in real time.

Stream Transcription puts live subtitles on top of any tab. The audio is transcribed (and translated, if you want) by YOUR OWN Whisper server — a free WhisperLive server on your computer or your network. Nothing goes to the cloud.

WHAT IT DOES
• Subtitles over any page: Twitch, YouTube, Kick, any video. They follow you to fullscreen, you can drag them anywhere, and choose 1–4 lines and the text size.
• Translation: Whisper's own translation to English, or ~100 languages with the server's translator.
• Any Whisper model, from the extension: paste a Hugging Face name or link (large-v3, turbo, a model fine-tuned for your language, a direct .zip…) and your server downloads it. Normal "transformers" models are converted automatically. Then just pick it from a list — no paths, no terminal.
• Made for real graphics cards: an "int8" option halves the model's memory with the same text, the translator can run on the CPU, and if a model doesn't fit you get a clear VRAM warning instead of subtitles that silently freeze.
• Save the subtitles as an .srt file.
• Interface in 15 languages.

YOU NEED A SERVER
This extension is the client. It needs a WhisperLive server (free, open source). The easiest way is the Stream Transcription server with Docker — step-by-step guide at https://github.com/javimentallab/stream-transcription#server. A plain WhisperLive server also works, with fewer features.

PRIVACY
The tab's audio is sent only to the server address you type. No account, no analytics, no tracking, nothing sent to the developer. Only when you press 'Check for updates', the add-on asks GitHub for the number of the latest version; nothing else is sent, and nothing is checked in the background. Open source (MIT).