BiteWise

Inspiration

We were university students struggling to find easy, budget-friendly meals. BiteWise was our solution to simplify meal planning while keeping it healthy and affordable.

What it does

BiteWise uses AI to help users create personalized meal plans. By entering their food preferences, favorite cuisines, dietary needs, and budget using simple prompts, users receive tailored meal prep recipes. They can then select recipes, generate a shopping list with all necessary ingredients, and get an estimated total cost for their groceries. Additionally, for every £30 spent, a tree is planted through the Verdn API.

How we built it

We used VS Code and utilized Ollama to prompt engineer our LLM. The project was developed using Node.js, CSS, Python, Flask, and MongoDB for database management. We integrated the Verdn API to facilitate tree planting for sustainable meal prep. Following software engineering best practices, we started with brainstorming, then moved through requirements gathering, design, wireframing, and finally coding.

Challenges we ran into

Sleep deprivation, coding failures, and training our LLM were major challenges. Integration was particularly difficult, as we were using 3 different operating systems which were incompatible, causing unexpected issues. Setting up and configuring MongoDB also proved to be a challenge, requiring troubleshooting to ensure seamless database operations.

Accomplishments that we're proud of

We successfully completed the entire system, creating a powerful tool that makes meal planning effortless. Our LLM provides highly accurate, reliable, and even entertaining recipe suggestions. The platform seamlessly integrates AI-driven meal planning with sustainability, helping users make informed, budget-conscious food choices while contributing to environmental efforts. With every purchase over £30, a tree is planted, making meal prep not just convenient but impactful. Additionally, we ensured that our solution remained concise and traceable to initial requirements, carefully distinguishing between essential features and unnecessary extras to maintain efficiency and focus. Another key achievement was successfully hosting MongoDB online, ensuring scalability and accessibility for our platform. Furthermore, our project management and design approach was highly effective, helping us outline the project's structure and ensuring a well-organised, visually appealing, and user-friendly interface.

What we learned

We proved to ourselves that we could develop a fully functional project in just 24 hours. We gained hands-on experience in training an LLM, fine-tuning it with LoRA, coding in Node.js, and working with MongoDB. Additionally, integrating and utilising the Verdn API in our code reinforced the importance of merging technology with sustainability. User feedback played a crucial role in refining the experience, and we learned the value of keeping processes simple and efficient.

What's next for BiteWise

We plan to introduce monthly meal plans, allowing users to plan their meals more effectively. Expanding to store user information will enable a more personalised experience, helping users track their preferences and dietary needs over time. Additionally, we aim to introduce a feature where users can find local stores for ingredient availability, ensuring convenience and accessibility in meal preparation. Our goal is to make meal prep smarter, easier, and greener.

Built With

• express.js
• llama3.2
• mongodb
• node.js
• ollama
• python