Econome

A FastAPI project that helps users create recipes from their available ingredients, reducing food waste and inspiring creative cooking.

Overview

Econome solves a problem I often face myself; not knowing what to cook with leftover ingredients. Using Mistral's SDK it generates recipes from both text descriptions and images of ingredients (e.g the inside of a fridge).

Quick Links

• API Documentation
• ReDoc Interface
• API Endpoints
• Setup Instructions
• Testing

Features

• Recipe generation from text input
• Recipe generation from images (using Mistral's Vision API)
• Recipe storage and management (CRUD operations)
• Optional recipe illustrations using FAL.ai

Implementation Details

• Python 3.13
• FastAPI
• Mistral SDK
• MongoDB (to enable faster prototyping)
• Modern UV package manager
• Extensive logging and exception handling

Getting Started

Prerequisites

• Docker
• Mistral API key
• FAL.ai API key (optional, for flux recipe illustrations)

Environment Setup

1. Create a .env file in the project root:

MISTRAL_KEY=<your-mistral-key>
FAL_KEY=<your-fal-key>  # Optional

2. Start the application from the top level:

> docker compose up

The server will be available at 0.0.0.0:8000

API Endpoints

Generate Recipe [POST /api/recipes]

Generates a recipe and optional illustration from a list of ingredients.

Request

[
    {
        "name": "sirloin steak",
        "quantity": "500g"
    },
    {
        "name": "bell peppers",
        "quantity": "2"
    },
    {
        "name": "soy sauce",
        "quantity": "500ml"
    },
    {
        "name": "oyster sauce",
        "quantity": "500ml"
    },
    {
        "name": "sesame oil",
        "quantity": "500ml"
    }
]

Response

{
    "meal": "Soy-Sesame Beef Stir-Fry",
    "steps": [
        {
            "content": "Cut the 500g of sirloin steak into thin strips, ensuring you're cutting against the grain for tenderness."
        },
        {
            "content": "Slice the 2 bell peppers into thin strips as well, then set them aside."
        },
        {
            "content": "In a bowl, mix together 250ml of soy sauce and 250ml of oyster sauce. This will be your sauce base."
        },
        {
            "content": "Heat a large pan or wok over medium-high heat and add 1 tablespoon of the 500ml of sesame oil."
        },
        {
            "content": "Once the oil is hot, add the beef strips and cook until browned and cooked through. Remove the beef from the pan and set it aside."
        },
        {
            "content": "In the same pan, add another tablespoon of sesame oil. Add the sliced bell peppers and cook until they begin to soften."
        },
        {
            "content": "Pour the soy and oyster sauce mixture into the pan with the peppers. Stir well to combine."
        },
        {
            "content": "Return the cooked beef to the pan and stir to coat everything evenly in the sauce."
        },
        {
            "content": "Cook for an additional 2-3 minutes to allow the flavors to meld together."
        },
        {
            "content": "Drizzle a small amount of sesame oil over the top before serving for added flavor."
        },
        {
            "content": "Serve hot with steamed rice or noodles."
        }
    ],
    "image_url": "https://fal.media/files/panda/T4KFDFAcIle1MAcGh6GHY_f5a4d26c3ca44e81b135171a389d48be.jpg"
}

Additional Endpoints

• Generate Recipe from Image [POST /api/recipes/vision]
  Generate a recipe from an image of ingredients

• Save Recipe [POST /api/recipes/save]
  Store a recipe in the database

• Delete Recipe [DELETE /api/recipes]
  Remove a recipe from storage

• Get Recipe [GET /api/recipes/{id}]
  Retrieve a specific recipe by ID

• List Recipes [GET /api/recipes/]
  Retrieve all saved recipes

Testing

For the best testing experience, you can access the interactive API documentation at:

• Swagger UI Documentation - Interactive API testing interface
• ReDoc Documentation - Alternative documentation viewer

To test the API:

1. Start the server using docker compose up
2. Navigate to http://localhost:8000/docs
3. Click on any endpoint to expand it
4. Click "Try it out" and input your test data
5. Execute the request and view the results