Innit : AI Sous-Chef for Personalized Cooking Guidance & Pantry Intelligence

Food Tech

Innit – Smart Cooking Assistant Powered by AI

Innit’s vision of transforming kitchens into intelligent spaces became reality with our AI-powered culinary assistant. We built a system that customizes recipes based on what users already have at home—thanks to pantry recognition and ingredient tracking. It offers real-time cooking instructions, smart substitutions, and voice-controlled guidance tailored to dietary needs. Whether it’s keto, vegan, or gluten-free, the AI adapts recipes on the fly. The assistant also optimizes cooking time and complexity for each user’s skill level. Innit is now more than an app—it’s a trusted AI sous-chef.

Project Overview

Client: Innit (U.S.-based food tech innovator redefining the connected kitchen)
Challenge: Home cooks struggled with meal planning, substitutions, and dietary adjustments—especially when working with what they already had at home
Goal: Build a smart culinary assistant to:
- Analyze pantry inventory and recommend real-time recipe matches
- Provide step-by-step cooking guidance tailored to dietary needs and skill levels
- Suggest smart ingredient swaps and cooking methods based on appliance usage
Team: 8 (3 ML Engineers, 2 Nutritionists, 2 Full Stack Developers, 1 PM)
Timeline: 4.5 months (Pantry AI Engine → Voice Assistant → Recipe Personalization Engine)

“With GenX’s AI, Innit isn’t just suggesting meals—it’s co-piloting the cooking journey. From what’s in the pantry to what’s on the plate, we’ve redefined home cooking.”

VP of Product Innovation, Innit

The Challenge

Critical Pain Points:

Users often lacked clarity on what they could make with existing ingredients
Dietary requirements made traditional recipe search rigid and irrelevant
Cooking skill levels varied widely, making instructions difficult to generalize

Technical Hurdles:

Building a unified food framework linking ingredients, cuisines, and possible substitutions.
Tracking pantry depletion without hardware sensors
Delivering real-time cooking assistance via voice and smart appliance integration

Tech Stack

Component	Technologies
Recipe NLP & Food Ontology	spaCy, BERT, Custom Ontology Framework
Personalization Engine	TensorFlow, Scikit-learn, NutritionIX API
Backend & APIs	Node.js, Firebase, PostgreSQL
Frontend Interface	React Native, WebRTC, GraphQL
Voice & Device Sync	Alexa SDK, Google Assistant SDK, MQTT
Analytics & Feedback	Segment, Mixpanel, Firebase Analytics

Key Innovations

AI adapted recipes based on available pantry items and user diet needs. Voice guidance, smart substitutions, and skill-based pacing enhanced usability. Every home cook got personalized, guided support—like having a sous-chef.

AI Recipe Adaptor

Rewrites recipes on-the-fly for diet, appliance, and user skill level

Result: 73% increase in recipe completion rates vs. static cooking apps

Pantry-Driven Meal Recommendations

Recommended dishes depending just on items consumers already owned.

Result: 61% reduction in mid-recipe grocery runs

Voice-Guided Smart Cooking

Real-time instructions adjusted by user pace and cooking environment

Result: 49% boost in daily active usage of the app’s voice mode

Our AI/ML Architecture

Core Models

Recipe Personalization Engine:
- NLP-based recipe parser that adjusts ingredients, portions, and cooking times
- Adapts instructions based on cooking skill, appliance availability, and dietary flags
Pantry Intelligence Module:
- AI identifies usable ingredients from scanned receipts, user inputs, and inventory tags
- Predicts what needs to be restocked and suggests meals based on what’s left
Smart Substitution & Guidance Engine:
- Suggests alternative ingredients and methods using culinary science + nutrition modeling
- Provides step-by-step instructions with voice or screen navigation

Data Pipeline

Sources
- Recipe databases, nutrition labels, user-uploaded pantry lists, grocery APIs
- User cooking history, dietary profiles, voice assistant logs