Affordable Data Summarizer
Get summaries of large text files and compare results from different AI providers
Affordable Data Summarizer is super simple to use - browse for your text file, or drag-and-drop it, then stand back:
In the example file above, I dropped in the text for another of my posts How UIPath helped me migrate my church's sermons to Spotify (substack.com) and waited for a few seconds. It’s nice to see they came back with different responses, so I have a couple of options to choose from.
Want to do this yourself?
First, I’ve spent a lot of time on this project and would appreciate any support you can provide, but feel free to try the project first and come back if you remember:
Instructions for running Affordable Data Summarizer
Create an OpenAI API Key
Create an Azure OpenAI resource and get the API Key by following How-to - Create a resource and deploy a model using Azure OpenAI Service
Clone the project repo:
git clone https://github.com/AffordableData/Summarizer
Rename the file .env-sample to .env
mv .env-sample .env
Update the values in the .env file with the information from steps 1 and 2 above.
Install the requirements:
pip install -r requirements.txt
Start the application:
streamlit run app.py
Open the website, which is usually
https://localhost:8501
if you haven’t run Streamlit apps before.
A word of caution
There is one more thing I want to mention - Cost!
Be careful, it’s easy to burn through a pile of cash by playing with temperature to increase or decrease the creativity of the summaries. For me, the transcript I used for testing was 10,002 tokens, and each time I created a summary was about like playing a video game back in the good old days - $0.26 … per provider! You’ve been warned!
Where next?
Going through the process of building this app has generated a ton of additional ideas that I want to explore! I’m sure I’ll be building more and writing about it again soon.