Our project is an AI-powered recipe builder, a feature extended from the application -- Kitchen Copilot, targeting leftover ingredients that many households often struggle to clear before the expiry date. This recipe builder is backed by Food Science, where we help users make better-informed meal planning decisions. With an image uploaded/taken, a few potential recipes, specifically in Asian Cuisine (or whatever cuisine they wish) will be AI-generated. This main feature will be complemented with further details of each product explained by Food Science (assisted with AI), e.g. how to extend shelf life of items, or how you can use these nearing end-of-shelf-life items in a creative manner as replacement for ingredients not found locally.
Our solution aims to target the lack of Singapore-contextualisation of the recipes found on Kitchen Copilot, as well as other prominent issues in Singpaore, such as food wastage. Our feature also has potential in adding a social aspect to the application with its extensibility, adding a special touch to Kitchen Copilot.
Video Demo: https://youtu.be/9Rj_Gpe3A3Q
- NodeJS version > 20
- Yarn package manager
- Navigate into frontend directory
cd frontend - Install yarn files (make sure you have npm and yarn installed)
yarn install - Run application
yarn start
- Docker
- Python version > 3.12
- OpenAPI account API key (can be obtained here https://platform.openai.com/settings/organization/api-keys)
cd backend- Setup python environment
python -m venv venvvenv\Scripts\activate - Install python libraries
python -m pip install -r supporting_files/requirements.txt - Put your OpenAPI key inside the .env.sample file, and remove the
.samplefile extension - Run backend server
python main.py - Run Ollama in a Docker Container
docker run --name ollama -d -p 11434:11434 ollama/ollama - Download LLaVA Inside the Container
docker exec -it ollama ollama pull llava - To stop container
docker container stop ollama
Make sure you are already inside the python venv then type
pip freeze > requirements.txt
Kah Ho - @kahkoii Dong Kiat - @dongkiat Jeya - @appdevin Tze Xuan - @tze-xuan