I wanted to use the opportunity of a Streamlit LLM Hackaton to build an app that integrates AssemblyAI for transcription, LlamaIndex for chatting with transcripts and Weaviate to store transcript embeddings in the Cloud.
In this video, you'll see me speedrun through my whole process, from
- downloading and extracting audio from my Youtube videos
- transcribing 2 years of Youtube content with AssemblyAI
- ingesting the transcripts with llamaindex + OpenAI
- storing and retrieving embeddings through Weaviate
- chatting with the transcripts in a Streamlit app
- detail features one could dive further into to improve the POC
After watching this video, you will have a good grasp on how to quickly setup a Streamlit Chatbot with a RAG (Retrieval Augmented Generation) process to generate answers to questions over Youtube videos.
00:00 Intro
01:06 Download audio
02:12 Transcribe audio
04:18 Vectorize audio
05:33 Store vectors
07:39 Chat with vectors
09:42 Areas of improvement
☕ Want to support me? https://www.buymeacoffee.com/andfanilo
🐦Follow my daily updates on Twitter: https://twitter.com/andfanilo
🗣️ Find me on: https://andfanilo.com/
👉 Links
- My Code: https://github.com/andfanilo/streamlit-chat-with-youtube-playlist
- Streamlit + Llamaindex : https://blog.streamlit.io/build-a-chatbot-with-custom-data-sources-powered-by-llamaindex/
- AssemblyAI: https://www.assemblyai.com/
- Llamaindex: https://www.llamaindex.ai/
- Weaviate: https://weaviate.io/
- Streamlit + Weaviate Magic The Gathering Search App: https://weaviate-magic-chat.streamlit.app/
My tools (Affiliate links to support me!)
- 📷 Main Camera - https://amzn.to/3QvTosB
- 📷 Main Lens - https://amzn.to/3svYCwq
- 🎙️ Microphone - https://amzn.to/47l5ewA
- 🎵 Music & Sound Effects - https://www.epidemicsound.com/referral/a2p22s
- 📕 Streamlit Official Book - https://amzn.to/3QuRPwl
______
🪶 It looks sponsored but I promise you it's not ahah!
⚠️ Disclaimer: This video is not sponsored, I receive no compensation by any brand quoted in this video. Views are my own and do not represent my employer's.
Links included in this description might be affiliate links. If you purchase a product or service with the links that I provide I may receive a small commission. Thank you for supporting my channel so I can continue providing you with free content!
#streamlit #python #datascience #dataapps
15 Comments