This content originally appeared on DEV Community and was authored by George Logue
Whiskey & Ember AI Bartender Teacher with Google AI Studio 
Introduction
At Whiskey & Ember, we believe cocktails are more than drinks—they’re stories told in flavour and fire.
For the Google AI Studio Challenge, we built the Whiskey & Ember AI Bartender Teacher, a tool that blends Gemini’s structured intelligence with our brand’s craft-driven style. The app takes ingredients you have on hand, generates creative cocktail recipes, and teaches you step by step how to prepare them—complete with substitutions and educational notes.
Live Demo: AI Bartender Teacher
Source Code: GitHub Repository
How We Used Google AI Studio
The innovation lies in prompt engineering with Gemini. Instead of free-form text, we required Gemini to return structured JSON objects with predictable fields. This allows our appUpload to render recipes consistently and highlight learning opportunities.
json
{
"cocktail_name": "Forest Whisper",
"ingredients": [
{"name": "Whiskey", "amount": "2 oz"},
{"name": "Honey Syrup", "amount": "0.5 oz"},
{"name": "Lemon Juice", "amount": "0.75 oz"}
],
"alternatives": {
"whiskey": "Canadian Rye or Bourbon",
"honey syrup": "Maple Syrup"
},
"steps": [
{"step_number": 1, "instruction": "Shake all ingredients with ice"},
{"step_number": 2, "instruction": "Double strain into chilled coupe glass"},
{"step_number": 3, "instruction": "Express lemon twist oils and garnish"}
],
"educational_notes": "This recipe demonstrates balancing sweetness and acidity using classic ratios."
}
This content originally appeared on DEV Community and was authored by George Logue