Inspiration

Many people buy groceries with the best intentions but end up wasting food because they’re unsure how to turn their ingredients into meals. I wanted to solve this by using AI to instantly turn a simple grocery receipt into smart, usable recipe suggestions making cooking easier, reducing food waste and making everyday life more convenient.

What it does

ReceiptChef lets users upload a photo of their grocery receipt, then uses OCR to extract a list of ingredients. Based on these, it generates personalized recipes using AI, taking into account factors like dietary preferences, time constraints and difficulty levels.

How we built it

This was built using Bolt.new. React + Tailwind CSS for the frontend Supabase for authentication, data storage, and saved recipes Tesseract.js for client-side OCR to extract text from receipt images

Challenges we ran into

OCR accuracy: Extracting clean, structured text from messy, faded, or crumpled receipts was a major challenge. Ingredient parsing: Not all grocery items are recipe-friendly as-is (e.g., brand names or vague descriptions), so we had to create rules to intelligently map them to real ingredients.

Accomplishments that we're proud of

Built a fully working end-to-end recipe engine from scratch. Enabled a smooth image-to-recipe pipeline entirely in-browser with a privacy-first approach. Designed a UI that feels like a mobile app but runs entirely in the browser with no install. Implemented smart ingredient mapping and recipe filtering in a scalable way.

What we learned

How to implement OCR in the browser and clean real-world text. Prompt engineering for real-world AI tasks like recipe generation. Building a seamless mobile-first experience using Tailwind CSS and React. Working with Supabase for full-stack functionality including auth, storage, and realtime data.

What's next for ReceiptChef

Smart meal planning: Let users create weekly plans and shopping lists based on receipts. Multi-receipt history: Track nutrition and food trends over time. Recipe rating system: Learn from user preferences to offer better recommendations. Localization: Support different languages and regional grocery formats. Pro version: Premium features like diet tracking, expert-curated recipes, and integration with health apps.

Built With

  • bolt.new
  • react
  • supabase
  • tesseract.js
Share this project:

Updates