This is a starting point for building a generative UI that searches the transcripts of CodeTV's video library. It uses OpenAI to create transcripts and embeddings for each video; ChromaDB to store the embeddings and perform vector search against them; and finally Hashbrown to create a generative user interface for the results.
nvm install&&npm install- Copy
example.envto.envand add your OpenAI API Key - Run ChromaDB using
npm run chroma - Download content.zip and unpack it in the root directory. This should create a folder called
content, with a subfolder calledtranscripts. I don't recommend opening up thetranscriptsfolder because of how large it is. - In a separate terminal, run
npm run push-transcripts-to-chroma. This should only take a few minutes. It is pushing all 140,000 or so transcript chunks to Chroma.
Now, with Chroma running, you should be able to launch the frontend and backend dev servers in separate terminals:
Frontend
npm run frontendBackend
npm run backend