This project allows anyone to upload a video file and be presented with a transcript and extracted questions & answers. The target user is a sales executive that wants to look back at questions asked during a call with a prospect/customer. Here is a link to a live demo if you don't want to deploy locally.
Must have git
installed. If you don't here is a great guide.
To run this project, you will need to add the following environment variables to your .env
file in the *same folder path as the app, summarizer.py
.
$ touch .env
$ vi .env
copy and paste your key/value using API_KEY=<your_api_key_no_quotes> format
:wq -> ENTER
Note: if you see a KeyError: 'upload_url'
if means you just need to paste in your API key appropriately.
To deploy this project run:
git clone https://github.com/ConnorBrereton/Sales-Insights.git
Next, naviate to the app directory.
cd Sales-Insights/
Install all of the dependencies using pip3
pip3 install -r requirements.txt
To run the application using Streamlit do the following:
python3 -m streamlit run summarizer.py
You should see this image pop up on localhost:8501
automatically.
Below is a demo of how the application can be utilized for call analysis to extract questions and answers with full context.
- Change color scheme to match AssemblyAI.
- Breakdown of speakers using Assembly's Speaker Diarization.
- Dockerize the application to avoid all dependency management.
- Remove filler words from Q&A - Breakdown.
- Use profanity filtering to see if prospect/rep used profanity during the call.
- Add webhooks to avoid using temporary variables to manage state within the application.
- Can use the ChatGPT API like in this project to extract Q&A faster and put it into a ordered form without the parsing algorithm.