Start with your ingredients
Enter ingredients and seasonings, choose a cuisine, and set servings and an optional cooking-time limit. Input validation catches empty entries and invalid values.
Independent project · Console application
What’s for Dinner?
Turn what’s in your kitchen into ideas for your next meal.
I built this C# application to suggest recipes from available ingredients and cooking preferences. It connects to OpenAI, validates the structured response, and presents the result in an interactive console.
Features
Find a practical suggestion
Prioritize available ingredients. If the AI reports no match, the app can make one additional request allowing up to three extra ingredients, clearly marking missing items.
See the recipe clearly
Read ingredient quantities, estimated preparation and cooking times, and step-by-step instructions for your requested servings.
Adjust without restarting
Request another recipe or edit individual preferences. Add, delete, or replace ingredients; duplicate additions are skipped. Apply changes or discard them without an API call.
From ingredients to a recipe
Set your preferences
Enter ingredients such as chicken, rice, broccoli, and mushrooms. Choose seasonings, cuisine, servings, and a cooking-time limit.
Explore a suggestion
Review the generated recipe and instructions. Choose “Try another recipe” to request an alternative using the same preferences.
Refine your ingredients
Choose “Edit my preferences” to update your list or cooking settings, then generate a recipe with the changes.
Behind the application
The application separates console interaction, recipe generation, data models, and validation. It uses built-in .NET HTTP and JSON libraries, with no external NuGet packages required.
- Console interface
- Collects and validates input, manages preference edits, and displays recipes and errors.
- API integration
- Sends cooking preferences to the OpenAI Responses API and requests output matching a JSON schema. The API key is read from an environment variable.
- Recipe validation
- Checks servings, times, required content, and ingredient references. It recalculates missing-ingredient status and rejects invented inventory references.
- Error handling
- Reports API failures such as invalid credentials and unavailable models. Errors do not trigger an automatic fallback request.
Development status
Working console application, demonstrated locally. The current version supports recipe generation, alternative suggestions, and preference editing during a session.
- Requires internet access and an OpenAI API key with API billing or credits.
- Preferences and recipes are not saved between runs; there is no persistent pantry.
- Tracks ingredient names, not available quantities. Cooking times are AI-generated estimates.
- Relies on AI suggestions rather than an exhaustive recipe database.
Read the setup instructions and current limitations on GitHub →