Inspiration
We noticed that many people struggle to find clear and reliable health information online. This inspired us to build something simple that can guide users through a conversation.
What it does
It’s a chatbot that understands user problems and gives helpful suggestions, mainly for health-related queries but also for general issues.
How we built it
We used Llama2 for responses, LangChain to manage the flow, Pinecone for storing data, and Flask to create a simple interface.
Challenges we ran into
Ensuring accurate responses and understanding user intent were the main challenges, along with integrating all components smoothly.
Accomplishments that we're proud of
We built a complete working chatbot that can understand queries and provide useful, easy-to-understand answers.
What we learned
We learned how to build real-world AI systems using LLMs, handle data effectively, and focus on user-friendly design.
What's next for MedBot
We plan to improve accuracy, add multiple languages, and make the chatbot more accessible for a wider audience.
Built With
- and
- flask-for-the-web-interface
- langchain
- llama2-(llm)
- natural-language-processing
- nlp-based
- pinecone-(vector-database)
- python
- retrieval
- techniques