RE:searcher is an AI-powered app designed to assist academics by enabling them to interact with their documents and generate mind maps seamlessly. Developed during the SAIS 24-hour hackathon, it won first place for its innovative approach to academic research.
This is the original repository, unchanged, as it was submitted on the competition. We continued developing the application; more about you can see here: https://re-searcher.web.app/
- Question formulation
- Creating summarized content overviews
- Efficiently saving user notes
The goal is to provide users with a simple way to search and manage documents, formulate questions, create notes, generate paragraphs, and define document topics and abstracts.
RE:SEARCHER - Complete documentation in Serbian
Our project repository consists of two directories: backend and frontend.
The frontend of our application uses Flutter for the user interface and is hosted on Firebase. For efficient state management, the BLOC pattern is employed. Thanks to Flutter, the application can run on all mobile and desktop platforms, as well as on the web.
The backend of our application uses the Flask micro-framework, enabling efficient interaction with the chatbot through message processing, document retrieval, and authentication. It leverages AWS services, the RAG model, and Pinecone vector DB for precise data search and management, ensuring relevant information and secure communication.
- Merging multiple documents
- Exporting notes to PDF format
- Grouping documents
- Image recognition
- Voice communication
- Navigating through document history
Introducing the team behind the project RE:searcher!
| Tara Pogančev | Vuk Čabrilo | Milan Popđurđev | Laslo Uri |
|---|---|---|---|
 |
 |
 |
 |
 |
 |
 |
 |