Summarize/Explain the doc

• 

  Original document

• 

  Summariing document

• 

  Application Frontend

Inspiration

Some of the documents are not easily understood by many. Some documents are long with difficult terms for a non-English speaking person or a student who is trying to learn the concept or a topic. Hence "Summarize/Explain doc" summarizes/explains and speaks the summarized/explained content in a way that is easy to understand using AI.

What it does

"Summarize/Explain docs" has a simple user interface where the user can upload a .pdf document. The document's content are shown in one section of the web browser. Once uploaded, the user can click on "summarize text" button and a summary is displayed in the other section of the web browser. The audio/sound button clicked once, plays the summarized text. If the sound/audio button is clicked twice, audio stops playing.

How it was built it

"Summarize/Explain doc' is built in Java Script, AWS Lambda, Open AI, Amplify and Vite. The front end is designed using Type Script / Java Script. Vite is used to create and package Java Script application, AWS Lambda is used to get the API keys from Open AI and encrypt it. The Lambda API Gateway transmits the encrypted keys and the keys are decrypted in the client-side. Once the keys are decrypted Open AI validates them, reads the documents and provides a summarized text for the the uploaded document. The packaged application is hosted using Amazon Amplify. The application is pushed to Github and each time a change is made, committed, and package is pushed to GitHub, the Amplify takes in the changes and it is deployed.

Challenges we ran into

Some of the challenges includes trying to convert the .pdf document into readable text, trying to get the right token count so that minimum three pages of the pdf document are uploaded. The right parameters need to be included in the Openai to take in most of the pdf pages and summarize. Also, trying to get the package deployed in Amazon's amplify took some effort as the right dependencies need to be included in the packaged json file.

Based on the application and resource consumption and availability the token can be increased to take in more text from the pages.

Built With

• amplify
• aws-lambda
• github
• javascript
• open-ai
• vite