Talk to any YouTube video
- Create a json file
tokens.json
that will have your model provider tokens in the following format
{
"openai": "<YOUR_OPENAI_TOKEN>",
"anthropic": "<YOUR_ANTHROPIC_TOKEN>",
"...": "..."
}
- Install the requirements:
pip install -r requirements.txt
- Run the server from the
talk-yt
directory:python websocket.py
- Open any Chromium-based browser (e.g. Google Chrome, Opera, Brave)
- Go to the extension page at
chrome://extensions
, or by following sandwich button > More Tools > Extensions - Make sure "Developer mode" is enabled, usually in the top-right corner
- Click "Load unpacked" and select the
chromium
directory from thefrontend
directory.