AI Party Card Generator

This is a submission for the Google AI Studio Multimodal Challenge

What I Built

I built the AI Party Card Generator, a web application designed to make creating personalized, high-quality birthday cards a fun, fast, and creative experience.

The core idea is to solve a common problem: you have a great, specific idea for a birthday card (like "a golden retriever astronaut celebrating on the moon"), but you lack the artistic skill or time to bring it to life. Stock photo sites are often too generic. The AI Party Card Generator bridges that gap. Users simply type a description, and the app leverages the power of Google's AI to instantly generate a gallery of ten unique, beautifully rendered cards.

But it doesn't stop there. The app also acts as a creative partner. It provides inspiring suggestions to overcome writer's block and, most importantly, allows users to select any generated card and edit it with simple text commands. This iterative process turns a simple generator into a powerful creative tool, ensuring the final card is absolutely perfect. To make the experience more magical and premium, each generated card features a subtle, shimmering animation, making the gallery feel alive and dynamic before the user even interacts with it.

Demo

Try it live here: LINK TO DEPLOYED APPLET

Here's a look at how it works:

1. Generate a Gallery of Ideas
A user enters a creative prompt, and the app generates 10 distinct card options in a beautiful, animated grid.

2. Get Inspired
If you're stuck, the app offers creative suggestions to get the ideas flowing.

3. Edit to Perfection
The real magic is in the editing. A user can select any card and make fine-tuned adjustments with natural language.

4. Share Your Creation
Once the card is perfect, it's easy to download or share directly with friends and family.

How I Used Google AI Studio

This applet is powered by the Gemini API, the same engine behind Google AI Studio. I integrated three distinct multimodal capabilities to create a seamless and powerful user flow:

1. Text-to-Image Generation (imagen-4.0-generate-001): This is the heart of the app. It takes the user's initial text prompt and generates the diverse gallery of ten high-quality birthday cards.

2. Image-and-Text-to-Image Editing (gemini-2.5-flash-image-preview): This is the "killer feature." To enable editing, I send the existing card image along with a new text prompt (e.g., "add 'Happy Birthday!' in the sky") to the Gemini 2.5 Flash Image Preview model. It intelligently interprets the request in the context of the image and returns a newly generated, edited version.

3. Text-to-Structured-Data (gemini-2.5-flash with JSON output): To power the "Need inspiration?" feature, I prompt the Gemini 2.5 Flash model to generate a list of creative ideas and ask for the response in a specific JSON format. This allows me to reliably parse the suggestions and display them as clickable buttons in the UI.

Multimodal Features

The combination of multimodal features is what makes this applet truly special and elevates the user experience beyond a simple image generator.

• From Idea to Gallery (Text-to-Image): Generating ten options instead of one is a crucial design choice. It gives the user a sense of abundance and creative discovery. They aren't just getting one interpretation of their idea; they're exploring a range of possibilities, which often sparks further creativity.

• Creative Conversation (Image-and-Text-to-Image Editing): The editing functionality is the most powerful multimodal feature. It transforms the interaction from a one-off command into a creative conversation with the AI. A generated image might be 90% perfect. Instead of discarding it and starting over, the user can say, "I love it, but can you change the background to blue?" This iterative refinement is intuitive, powerful, and deeply engaging. It puts the user firmly in the director's chair, using AI as their personal artist.

By weaving these capabilities together, the AI Party Card Generator becomes more than a tool—it's a creative co-pilot, making personalized digital art accessible and fun for everyone.