Inspiration 💡🍳🥘
- This pandemic season, cooking was one of the few fun things to do around the house.
- Being a huge foodie, I got into the habit of cooking the meals I could no longer go out to eat, and started spending more and more time in the kitchen.
- The only issue is juggling a hot pan, sharp knives and your phone is a recipe for disaster (hehe get it?).
- Which is why I decided to build Michelin, an AI powered cooking assistant that allows you to navigate through recipe instructions only using your voice, completely hands free!
What it does ❓🤖🔊
Michelin helps you go hands-free and stress-free in the kitchen, by being your very own personal assistant. Pick any recipe you want, and Michelin will walk you through each ingredient to get your mise en place in place (the puns don't stop!), and walk you through each step of the recipe.
- Didn't catch a step? Ask it to repeat.
- Not sure if you did one of the previous steps right? Ask it to go to that step.
- Cooked this recipe a million times already? Ask it to skip to a future step, or skip to the end entirely.
- Burnt everything on the stove and need to start from scratch? Just ask.
How I built it 🔧🔨🧰
- The app fetches the desired recipe from AllRecipes.com, and has Dasha AI uses its excellent TTS to help the user navigate.
- I used Cheerio and Node-fetch to scrape the contents of the recipes, which allrecipes.com very kindly provided in an easily-scrapeable format.
- I used AWS to deploy the app to test, then promptly took it down out of fear of Bezos' wrath in the form of a $20,000 bill. 😂
- I used Dasha AI to set up an incredibly cool conversational AI interface to interact with the user.
- Plenty of coffee ☕ and not enough sleep 💤 😂
Challenges I ran into 🏃♂️🛑🚧
- Mostly figuring out Dasha's DSL.
- The app is simple enough to build out in JavaScript but DSL is not something I'm used to using.
Accomplishments that I'm proud of 🏆🏅
- Getting it to work.
- Learning a new skill (working with intents, entities, and a new stack).
- Wrapping up an app that works in the span of a weekend.
- Getting to use a new super-cool conversational AI app.
- Meeting cool people. The engineers and advocates from Dasha were super cool, and I look forward to staying in touch and building with Dasha.
What I learned 🧠
- DSL and working with Conversational AI.
- Webscraping using Cheerio and Node-fetch (traditionally I'd use BS4 and requests or axios).
What's next for Michelin ⏭
- Automatically set timers for baking.
- Get Michelin to tell you what you can cook based on what you have in your pantry.
- Order needed ingredients using e-commerce sites, and keep track of your pantry.
- Who knows? Maybe at some point, Michelin will do all the cooking for you! 😉
Built With
- amazon-web-services
- cheerio
- css3
- dasha
- dasha-ai
- dsl
- html5
- javascript
- node-fetch
- node.js

Log in or sign up for Devpost to join the conversation.