Added run.sh to help with running locally

Web generator outputs full transcript at the bottom, and highlights which chunk is playing

Support for OpenAI and the GPT-5 model

Improve handling of timestamps

Add skip argument to skip to a stage in the processing

Make sure the generated video summary refers to the video file in the same folder

Generate web page in output folder, then copy to destination

Streamline install, bypassing version check and directly fetch insanely-fast-whisper

device-id config, allow auto config with detection or manual override

Clean-up install instructions

Add the generated sample summary.

chore: add MIT license with Steampunk Digital Corp. copyright

test: fix test setup and mark integration tests as xfail

Revised the intermediary paths, and tests to use the test_data

Updated key_moments AI model to use claude-3-5-sonnet-20241022