recipies.ai

recipes.ai
Colorblindness Contrast Test

Inspiration

A certain degree of laziness spinkled with the practicality of removing blocks slowing people down in their busy days. We hope to no longer have the problem of looking through our pantries and fridges every five minutes, looking for something not found last time.

What it does

recipes.ai gets a uri image string from the user and returns 3 lists of recipes they could make with the ingredients provided (first list with an unmodified search, the second list with a search modified for a vegan diet, and the third list's search is modified for a vegetarian diet)

How we built it

Once a uri string is searched the string is passed to the Google Vision API, detecting objects and labels, the ingredients found are then passed to a culinary search engine, with the vegan and vegetarian searches being modified to specifically include those results, the resulting links are organized by popularity and displayed in the table (1 is the most popular, 7 is the least), all ingredients found in the photo are listed above the table. If the user wants to enter another search all they have to do is hit the reset button in the top-left corner and their page will reset to its default state.

Challenges we ran into

No matter what library or method we used, we were unable to find anything that would allow the website to directly convert image files into uri-compatible strings.

Accomplishments that we're proud of

Despite our inexperience with Hackathon and machine learning we were able to create a website capable of harnessing a machine learning algorithm created by Google for decreasing the amount of time people spend uselessly, allowing them to do something more productive with something as precious as time.

What we learned

We learned what to expect from beginner-accessible Hackathons, how to interact with more complex Python than we have before, and ways to make websites more accessible for users needing a modified web experience.