Currently the application in the repository supports only AI based image search which uses OpenAI's CLIP model to retrieve related images and finds most similar messages with the given prompt and proposes answer based on the similar message and question.
The application can be further redeveloped using
- Searching inside video and audio data
- Question answering based on images, etc.
Open to collaboration and pull requests.
See how application works: