ChefAssist AI: WhatsApp Conversational Cooking Assistant
ChefAssist AI is an AI-powered WhatsApp chatbot platform that enables users to receive personalized cooking guidance through conversational interactions. Users scan a QR code on product packaging, access a landing page, and start a WhatsApp chat where the AI generates recipes based on ingredients, dietary preferences, and cooking constraints, with support for voice inputs and automated cooking timers.
Key Features
-
WhatsApp-Based Conversational Cooking Assistant
- Built a seamless WhatsApp-first user experience that allows users to interact with the cooking assistant through natural chat conversations.
-
AI-Powered Personalized Recipe Generation
- Implemented AI workflows that generate custom recipe suggestions based on available ingredients, dietary restrictions, and cooking preferences.
-
Multi-Model AI Orchestration
- Integrated OpenAI, LangChain/LangGraph, Groq, and Google Gemini to power intelligent conversational flows, context handling, and response generation.
-
Voice Input Processing
- Enabled users to send voice messages and leveraged AI-based speech-to-text processing to convert spoken queries into structured cooking requests.
-
Automated Cooking Timers and Guided Assistance
- Developed background workflows to manage timers, delayed prompts, and guided cooking steps, improving the hands-free cooking experience.
-
Webhook-Based WhatsApp Messaging Infrastructure
- Built a reliable webhook architecture to handle incoming messages, delivery flows, retries, and delayed processing across WhatsApp Business API.
-
Scalable Async Task Processing
- Used Celery and Redis to manage asynchronous jobs such as timer execution, delayed notifications, and background conversational tasks.
-
Interactive Packaging-to-Chat Experience
- Transformed physical product packaging into a digital AI engagement journey through QR-based access and instant conversational onboarding.
Challenges
-
Seamless WhatsApp Integration
- Handling reliable message processing and delivery through WhatsApp Business API webhooks required managing retries, delayed messages, and webhook reliability.
-
Multi-Model AI Response Pipeline
- Generating accurate conversational responses required orchestrating multiple AI models and managing prompt workflows efficiently.
-
Handling Voice Input
- Users needed the ability to interact with the chatbot using voice messages instead of only text.
-
Managing Asynchronous Tasks
- Features such as cooking timers, delayed notifications, and background message processing required a scalable async processing system.
Solution
-
Webhook-Based Messaging Architecture
- Integrated the WhatsApp Business API using a webhook-based architecture that reliably receives, processes, and responds to incoming user messages.
-
AI Conversational Workflow
- Implemented a conversational pipeline using OpenAI, LangChain/LangGraph, Groq, and Google Gemini to generate intelligent responses and maintain conversational context.
-
AI Voice Processing
- Integrated Whisper speech-to-text capabilities to process voice messages and convert them into structured queries for the AI recipe system.
-
Scalable Background Processing
- Implemented Celery and Redis to manage asynchronous workflows including timers, notifications, and background message processing.
