This is a program that I wrote to convert the ISW's "Russian Offensive Campaign Assessment" reports to an audio format I could listen to while working. It's very hacky, only supports these specific articles at the moment, and it relies on OpenAI's TTS-1-HD and GPT-4o models. It generates a summary of the main efforts and scrapes the provided summary of daily events, and it reads them out before the full report.
git clone https://github.com/acureau/isw-ai-podcast.git && cd isw-ai-podcastpip install -r requirements.txt- Install ffmpeg.
- Insert your API key into the
constants.pyfile. - Update the report URL in the
main.pyfile. - Execute
main.py, wait for the output.wavfile.
In the rare chance anyone wants to contribute I'd like to expand support to other ISW reports. If you fix a bug I will merge. If you're more ambitious the code is currently scraping image URLs, the original plan was to generate a slideshow and display those images when relevant sections were being read. I have no immediate plans to do this myself.