ChatGPT TTS Grabber

Automatically downloads ChatGPT's TTS audio whenever playback is triggered.

作者
dudebot
今日安裝
0
安裝總數
9
評價
0 0 0
版本
1.0.0
建立日期
2025-05-22
更新日期
2025-05-23
尺寸
1.5 KB
授權條款
MIT
腳本執行於

ChatGPT TTS Grabber

A Tampermonkey/Greasemonkey userscript that automatically captures and downloads the text-to-speech (TTS) audio generated by ChatGPT whenever you click the voice playback button.

✨ Features

  • 🔊 Automatically detects and downloads audio streamed via ChatGPT’s voice feature (including Glimmer and other voices)
  • 📥 Captures streamed audio via Media Source Extensions (MSE), even if traditional fetch hooks fail
  • 🐒 Works with Tampermonkey or any UserScript-compatible extension
  • 💡 Compatible with both chat.openai.com and chatgpt.com

📦 Installation

  1. Install Tampermonkey (or similar userscript manager)
  2. Install the script via GreasyFork (link coming soon) or from this GitHub repo
  3. Visit ChatGPT and use the voice playback feature as usual
  4. Watch your browser download the audio automatically as .webm, .mp3, or .aac files

⚙️ How It Works

ChatGPT uses Media Source Extensions (MSE) to stream TTS audio to its player. This script monkey-patches the browser’s MediaSource API to capture and reconstruct all audio chunks, packaging them into a downloadable Blob.

This approach works even if the audio is streamed directly and not returned as a traditional file from the server.

🧪 Known Limitations

  • May not work if ChatGPT changes its audio streaming pipeline to an AudioWorklet or encrypted transport
  • Audio is auto-downloaded for every playback (no opt-out UI yet)

🛠 Contributing

Feel free to fork or submit issues and improvements on GitHub.

📄 License

MIT