Chemical Buddy

• 

  Dashboard with a option to start conversation and a Chemistry Unit Converter

• 

  Some basic information about chemistry in dashboard

• 

  Option to upload pdfs and ask doubts

• 

  Answering the question with visuals

Inspiration

The inspiration for "Chemical Buddy" came from my own struggles with understanding chemistry concepts during my studies. We noticed that many students find it tough to understand things like chemical reactions and how to convert units.

What it does

It is an AI-powered chemistry study assistant designed to enhance the learning experience for users. It allows users to ask chemistry-related questions and receive detailed answers along with images related to their queries, providing valuable visual aids. Users can also upload PDF documents to extract relevant text, which aids in answering their queries. The application includes a unit converter for common chemistry conversions, such as moles to grams and molarity calculations.

How we built it

The front end was developed with Streamlit, providing a user-friendly interface. To process and manage natural language queries, we used Langchain for text splitting and conversational chains, and integrated Google Generative AI to enhance text understanding and generate informative responses. For document handling, FAISS was implemented to efficiently store and search text embeddings, while PyPDF2 was used to extract text from user-uploaded PDFs, allowing users to incorporate their own study materials.

Challenges we ran into

We faced challenges while extracting text from PDFs was tricky because of different formats, which sometimes caused the text to come out incomplete or jumbled. And while integrating the Google Custom Search API for fetching images, especially with making sure the results were relevant and working within the API limits.

Accomplishments that we're proud of

We successfully developed an AI chatbot that provides accurate, context-rich answers to chemistry questions. We also implemented a smooth document upload feature that processes user PDFs, expanding the knowledge base the assistant can use. The user-friendly interface encourages learning and makes chemistry concepts more accessible. Additionally, we integrated an image search feature, allowing users to access visual aids that enhance their understanding of chemistry topics.

What we learned

We gained experience in integrating multiple APIs and AI tools to create a unified application. Additionally, we deepened our understanding of natural language processing challenges, particularly in specialized areas like chemistry, and developed strategies for optimizing data storage and retrieval to enhance response times and overall user experience.

What's next for Chemical Buddy

In the future, we aim to introduce interactive learning features like reaction balancer, quizzes, flashcards, and experiments to engage users more actively. Additionally, we are considering developing a mobile version for greater accessibility and implementing personalized user profiles to tailor learning experiences to individual needs.

Built With

• faiss
• googlecustomsearchapi
• googlegenerativeai
• langchain
• pypdf2
• python
• streamlit