PDF Question Answering (QA) powered by ChatGPT.
A handy Chrome extension I built that leverages the power of Large Language Models (LLMs), such as ChatGPT, to ingest new data in the form of PDF documents and enable the user to perform QA tasks.
The project has two main components:
- Chrome extension frontend built using Flutter and Dart
- Application backend built using 🦜️🔗 LangChain, OpenAI's ChatGPT, and Google Cloud Functions for serverless deployment
To deploy and run the application yourself, you will need an OpenAI account with API key for accessing ChatGPT and a Google Cloud account to access the Cloud Functions service.
- Install Flutter: https://docs.flutter.dev/get-started/install
- Create a new Google Cloud Function with the source code found in
backend/
. You will need to provide your OpenAI API key as an environment variable namedAPI_KEY
- Make note of the deployed URL and update the
baseUrl
andurlExt
variables inlib/summary_api_client.dart
- Build the Chrome extension package with
flutter build web --web-renderer html --csp
- Load the build
build/web
into Chrome Extensions developer mode - Test out IntelliQAi on any PDF!