A simple youtube summarizer using a local AI ollama server. Today we have so much content online and so few time to be able to watch everything, that I've thought it would be cool to make a simple script to having a summary of the video previously to viewing it so I can decide if its worthy or not to see it.
Remember that this relies on having already a running Ollama Server.
1. Clone the repo
git clone https://github.com/HariTrigger/OllamaYTSumm/
2. Install the requirements
cd OllamaYTSumm
pip install -r requirements.txt
python3 main.py
- Update the code to adapt to changes made to youtube-transcript-api (several important changes made on v1.0.0)
- Work the system prompt to get the best result possible with (hopefully) the most ammount of models.
- Make the script asynchronous so we can activate the Stream of the response from the server, and see it real time.
- Make a web UI so I don't deppend on the console to do this. (Didn't tought on how yet, suggestions are welcome)
This is a script I've made for fun and to solve my own problem while learning through it.
Any contribution are welcome, and thanks in advance to the ones who do so... Thank you to the people that already did.